JPH0783817A - Flow-type particle image analysis method and flow-type particle image analysis device - Google Patents

Abstract

(57) [Summary] [Purpose] In a particle image analyzer that takes a static image of particles suspended in a flowing liquid and analyzes the particles, the number of particles in one image detected by a particle detection system and the static image Analysis accuracy is improved by associating with existing particles. A sample is flown into a flow cell 100, passage of particles is detected by a laser beam 10, a pulse lamp 1 is caused to emit light based on the detection, and a still image of the particles is picked up by a TV camera 8 to perform image processing. In the classification result totaling unit 70 of the particle analysis unit 103, the number of particles detected by the particle detection unit 102 during one image time, and the particle analysis unit 10
The particles in the sample are classified and the particles are counted by associating with the particle image in which the particles are classified by the pattern recognition processing in one image processing in 3.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a flow-type particle image analysis method and a flow-type particle image analysis apparatus, and more particularly to a particle image analysis for capturing a static particle image suspended in a flowing liquid and analyzing the particle. It is a thing. In particular, it is used to analyze cells and particles in blood or urine.

[0002]

2. Description of the Related Art In conventional particle image analysis, cells in blood and cells and particles in urine are classified and analyzed by preparing a sample on a slide glass and observing with a microscope. . In the case of urine, since the particle concentration in urine is low, the sample is observed after being centrifugally concentrated by a centrifugal separator.

As a method or apparatus for automating these observation and inspection operations, a sample such as blood is smeared on a slide glass and then set on a microscope, and the microscope stage is automatically scanned to detect the presence of particles. At this position, the microscope stage is stopped and a still particle image is captured, and the feature extraction and pattern recognition techniques of image processing technology are used to classify and analyze the particles in the sample.

However, in the above method, it takes time to prepare a sample, and it is necessary to find particles while mechanically moving the microscope stage and move the particles to an appropriate image capturing area. Therefore, there are disadvantages that analysis takes time and the mechanical mechanism becomes complicated.

In a particle image analysis method or a particle image analysis apparatus which does not form a smear as described above, a flow site in which a sample sample is suspended in a liquid and allowed to flow in a flow cell for optical analysis. The meter method is known.
The method using the flow cytometer observes the fluorescence intensity and scattered light intensity from each particle in the sample, and has a processing capacity of several thousand per second. However, it is difficult to observe the feature quantity that reflects the morphological characteristics of the particles, and the particles cannot be classified by the morphological characteristics that have been performed under the conventional microscope.

As an attempt to capture static particle images in a continuously flowing sample sample and classify and analyze the particles from each static particle image, Japanese Patent Publication No. 57-50099.
5, JP-A-63-94156, JP-A-4-
The technique described in Japanese Patent No. 72544 is known. In the technique described in Japanese Patent Publication No. 57-500995, a sample sample is passed through a channel having a special shape into a wide imaging area, and a static particle image is captured by a flash lamp.
A method for particle analysis using the image is shown.

In the method, when a magnified image of sample particles is projected on a CCD camera using a microscope, the flash lamp, which is a pulse light source, periodically emits light in synchronization with the operation of the CCD camera. Since the light emission time of the pulsed light source is short, a still image can be obtained even when particles are continuously flowing, and the CCD camera can capture 30 still images per second.

In the technique described in Japanese Patent Laid-Open No. 63-94156, a particle detection system is provided upstream of the particle image capturing area in the sample flow, in addition to the stationary particle image capturing system.
This is a method of knowing particle passage in advance by a particle detection system and turning on a flash lamp, which is a pulse light source, at an appropriate timing when the particle reaches the particle image capturing area.

In this method, it is possible to detect the passage of particles and take a static particle image at a timing only at that time without periodically emitting light from the pulsed light source. Is collected, and even in the case of a sample sample having a low concentration, a meaningless image in which particles are not present is not captured and processed.

Incidentally, as an example in which an ordinary flow cytometer or a particle analysis device which does not take a static particle image, has a particle detection system separately, JP-A-60-260830 and JP-A-63-231244 are disclosed. JP-A 1-2
The technology described in Japanese Patent No. 45131, etc. is known.

[0011]

In the above-mentioned conventional particle image analyzer, generally, a still image of continuously flowing sample particles is analyzed to determine the number and classification of plural kinds of particles in the sample. In order to perform efficiently, a particle detection system for detecting passing particles is required in the static particle image capturing area or upstream thereof, as is done in the above-mentioned known example. That is, the pulse light source is turned on only when the sample particles pass, and a still image of the sample particles is captured.

Since the above method can process a measurement sample having a low particle concentration very efficiently, it is possible to increase the amount of measurement sample, shorten the analysis time, and improve the analysis accuracy. However, as described above, in order to analyze or classify the number of plural kinds of particles contained in the sample by analyzing the still image of each particle which is continuously flowing, it is described below. There are such problems.

[1] According to the technique described in Japanese Patent Laid-Open No. 63-94156, in a flow-type particle image analysis device, a particle detection system for detecting passing particles in a still particle image capturing area or upstream thereof, and a particle detection system An image pickup system for picking up a still particle image by turning on the pulsed light source only when passing through is provided, but the particle detection system and the image pickup system generally have different measurement principles.

As the particle detection system, there is used a method in which a laser beam is focused on a measurement sample that normally flows in a flow cell and irradiated, and light scattered light from particles that cross the laser beam is detected. . The light scattered light generally produces an optical signal whose intensity is proportional to the effective scattering cross section of the particle.

The optical signal is converted into an electric signal by a photodetector. The magnitude of the optical signal is affected by the optical refractive index of the particles, absorption, size, the internal state of the particles, scattered light detection conditions, etc., the particle diameter is necessarily the morphological information of the particles used in the image processing, It does not completely match the perimeter, color information, and texture information.

When the particle detection system is set to detect particles at a predetermined particle detection level or higher, the size of an actual stationary particle image obtained by the image pickup system is small and particles which do not reach the particle detection level due to light scattering. Is present
It often happens. These relationships will be specifically described with reference to FIG. FIG. 6 is a schematic explanatory view of particle detection and still particle image capturing in the related art.
FIG. 6 (1) is a diagram showing a particle detection signal by passage of particles in a particle detection system, and the measurement principle of the particle detection is to detect the light scattering signal from the particles by the laser light as described above. And

In the diagram of FIG. 6 (1), the particles a and d are larger than a predetermined particle detection level. First, the particle a is detected by the particle detection system, and after a certain time elapses, it is stopped by the image pickup system. A flash lamp emits light to capture a particle image. FIG. 6C shows the captured still particle image.

In the picked-up image, a total of five a,
Particles b, c, d, and e exist, but when this captured image is made to correspond to the particle detection signal of FIG.
Since the two particles a and d are larger than the particle detection level, they are counted as the signal of the particle detection system, but the remaining c,
No particles are detected in d and e. The undetected particles are particles that do not need to be analyzed. The particle detection pulse signal by the particle detection system is shown in FIG. 6 (2).

By reducing the particle detection level, it is possible to detect all of these five a, b, c, d, and e particles. However, in that case, the frequency of particle detection increases significantly, and conversely, the detection rate of particles to be analyzed decreases. Further, the dead time due to the image capturing may cause counting down of particles. As a result, the particle analysis accuracy and reproducibility are deteriorated.

Further, generally, the number of particles detected by the particle detection system does not match the number of particles processed by the image pickup system. When there are many types of particles contained in the sample, the abundance ratio of the particles detected by the particle detection system,
This causes the abundance ratios of particles processed in the image pickup system to not match. This is because particles below the detection level of the particle detection system are often included as particles to be processed by the image pickup system.

In such a state, from the information on the total number of particles detected by the particle detection system and the information on the existence ratio with respect to all the particles processed by the image pickup system, the number of particles of each kind in the sample and the particle existence ratio are obtained. , It is not possible to know the particle density and concentration information correctly. Also, among the image-processed particles,
There may be particles that are smaller than the particle detection level of the particle detection system and that are the processing target of the image capturing system, or particles that are not the processing target particles but that provide important information.

If these particles are also analyzed correctly, how much they exist in the sample are analyzed, and if they are not reflected in the analysis result as particles to be processed by the image pickup system, the number of particles of each kind in the sample It is not possible to correctly know information such as particle existence ratio, particle density, and concentration.

Further, in the technique described in Japanese Patent Laid-Open No. 4-72544, a counter for counting all passing particles is used in the particle detection system. In this case, the number of particles of each type in the sample can be known from the existence ratio of the value of the particle counter and the number of particles subjected to image processing in the image pickup system.

However, for this purpose, the prerequisite is that the distribution of particles to be detected by the counter of the particle detection system and the distribution of particles to be image-processed by the image pickup system match. Is. However, as described above, in the particle detection system measurement and the particle processing by the image processing of the image capturing system, the measurement methods and principles are different, and the distributions by type often do not match.

The present invention has been made to solve the above-mentioned problems of the prior art, and is provided with a particle detection system and an image pickup system, respectively. The principle of the particle detection system and the particle image discrimination of the image pickup system are provided. In a flow-type particle image analysis method and a flow-type particle image analysis device having different principles, the detected particles and the particles in the still particle image are made to correspond to each other, and particles other than the detection particles of the particle detection system are static particle images of the image pickup system. Even if it exists in the sample, the flow type particle image analysis method and the flow type particle image analysis device can accurately know the number of particles of each type in the sample, the particle existence ratio, the particle density, and the concentration information, and the analysis accuracy is good. Is provided.

[0026]

In order to achieve the above object, the structure of the present invention relating to a flow type particle image analysis method is as follows:
A step of flowing a liquid sample to be suspended in a flow cell and counting the number of particles in the sample by a particle detection system; and a step of imaging a still image of particles passing through an imaging region in the flow cell by a particle imaging system. In the flow-type particle image analysis method, which comprises an analysis step of performing morphological classification of particles in the image by image analysis means, the analysis step includes one image particle count value by the particle detection system,
A particle distribution or particle classification in the liquid sample based on an extraction step in which each particle in the stationary particle image is associated with a particle image identification feature amount, and the corresponding relationship and the total particle count value by the particle detection system in the detection step. One of the steps is a particle analysis step.

The particle analysis step is characterized in that it has a plurality of measurement modes for the same liquid sample, analyzes each measurement mode, and integrates all data after completion of analysis in all measurement modes. The extraction step is characterized in that, as the particle image identification feature amount, any one of size information, color information, light scattering information, fluorescence intensity information of the particles, or information combining a plurality of these is used.

The particle analysis step calculates either the number of particles or the ratio per sample volume for each of the image processed images, and then uses either the number of particles or the ratio of particles to determine the particles per total sample volume. It is characterized in that the number is obtained and based on the number of particles. The particle analysis step is characterized in that either the number of particles in the measurement sample, the particle concentration per unit volume in the measurement sample, or the number of particles in terms of a microscopic field is obtained, and analysis is performed based on these.

In the particle analysis step, the remaining particles in the still particle image that could not correspond to one image particle count value of the particle detection system are analyzed as a separate frame, and at the end of the analysis, the particles corresponding to one image particle count value are analyzed. It is characterized in that the additional frame processing is added to the result. The particle analysis step is characterized in that processing is determined when the number of processed images is small and there is not a sufficient volume to be processed to perform another frame processing. The particle analysis step is characterized in that when the number of processed images is small and there is not a sufficient volume to be processed to perform another frame processing, the evaluation procedure is input and set beforehand.

The particle analysis step is characterized in that if the number of processed images is small and there is not enough processing target volume to perform another frame processing, another frame processing is not performed. The particle analysis step is characterized in that when the number of processed images is small and the processing target volume that is sufficient to perform the separate frame processing is small, the separate frame processing is performed and a problem in the processing is output.

A particle detection system for flowing a suspended liquid sample into a flow cell to count particles in the sample, a particle imaging system for capturing a static image of particles passing through an imaging region in the flow cell, and the stationary In a flow type particle image analysis device comprising an analysis means for performing morphological classification of particles of a particle image, the analysis means, one image particle count value of the particle detection system, and each particle in the stationary particle image And a particle analysis unit that performs either particle distribution or particle classification in the sample from the correspondence relationship and the total number of particles counted by the particle detection system. It is characterized by having done.

The particle analysis unit has a plurality of measurement modes for the same measurement sample, analyzes each measurement mode, and is configured to integrate all data after the analysis of all measurement modes is completed. To do. The extraction unit, as the particle image identification feature amount, either particle size information, or light scattering information, or color information, or fluorescence intensity information,
Alternatively, it is characterized in that the information is combined by using information obtained by combining a plurality of these.

The extraction unit is characterized in that only the particles corresponding to one image particle count value are made to correspond from the magnitude relationship between the one image particle count value and the particle morphological feature amount. The particle analysis unit is configured to calculate the number of particles or particle ratio per sample volume corresponding to all the image-processed images, and further to determine the number of particles per total sample volume. The particle analysis unit is configured to calculate the number of particles in the measurement sample, the particle concentration per unit volume in the measurement sample, or the number of particles in terms of a microscopic field.

The particle analysis unit analyzes the remaining particles in the still particle image that could not correspond to one image particle count value of the particle detection system as a separate frame, and at the end of the analysis, analyzes the particles corresponding to one image particle count value. It is characterized in that the result is added with the additional frame processing. The particle analysis unit is configured to perform another frame processing when the number of processed images is small and there is not enough processing target volume to perform another frame processing.

The particle analysis unit is characterized in that when the number of processed images is small and there is not a sufficient volume to be processed for another frame processing, the evaluation procedure is input and set in advance. The particle analysis unit is characterized by not performing separate frame processing when the number of processed images is small and there is not enough processing target volume to perform separate frame processing. The particle analysis unit, when the number of processed images is small and the volume to be processed is small enough to perform another frame processing,
It is characterized in that it is configured to display or output that there is a problem in another frame processing.

The particle detection system is characterized in that it can correct an erroneous count due to pulsed light source emission for still image capturing with respect to one image particle count value and the total particle count value. The analysis target is a biological cell.
The analysis target is a blood cell component existing in blood. It is characterized in that the object of analysis is urine, particularly urinary sediment components present in urine.

A more specific description will be given. Particles that have passed through the particle detection area in the flow cell are detected, and when the detected particles arrive at a predetermined position in the still image capturing area after a certain period of time, a flash lamp is turned on to capture a still image of the particle. At this time, the total number of particles having a level equal to or higher than a certain detection signal level existing in the portion of the sample flow captured in this one still image is defined as one image particle count value.

In general, in one captured still particle image, there are more particles than the one image particle count value of the particle detection system. This tendency increases with increasing particle detection signal level. Here, all the particles existing in one static particle image are classified and analyzed by using the particle identification feature amount that is the same as or close to the detection principle of the particle detection system.

As the particle image identification characteristic amount, it is possible to use morphological information or color information of measurement particles, fluorescence intensity information, or a characteristic amount obtained by combining a plurality of these information items. Specifically, from the particle image, morphological information such as particle size, area, diameter, perimeter, and projection image, particle concentration, color information, particle internal structure, and fluorescence intensity are used. Information such as the existence of nuclei, cytoplasm, intracellular granules, etc., like biological cells can also be used.

Further, the transit time of the particles through the particle detection system may be used as a parameter for the size of the particles. That is, it is checked which of the particles present in the image and which are the detection particles detected by the particle detection system. The number of particles that can be taken corresponds to one image particle count value. This is called regular identification processing. The remaining particles for which the detected particles and the particles in the image cannot be associated with each other are subjected to identification processing different from the corresponding particles. This process is named another frame identification process.

Furthermore, the count value of all particles passing through the particle detection system and above the detection level is measured regardless of the still image pickup by particle detection. In this case, it is necessary to prevent erroneous counting of particle detection signals due to flash lamp light. To prevent this, either the flashlamp light does not reach the particle detector or the flashlamp signal is not electrically counted. In addition, it is possible to prevent erroneous counting by detecting the flash lamp light and subtracting the number of times of lamp light emission at the end of measurement.

When processing dense particle samples,
Due to the dead time associated with image reading, particle counts may be missed in the particle detection system.Therefore, the total number of particles in one image particle count value during a certain measurement time does not always pass through the particle detection system during the certain measurement time. It does not show total particle counts above the detected level. For this, the total number of particles above the detection level in the sample within a fixed measurement time,
Count all particle detection counts.

Further, since the particles to be subjected to the separate frame discrimination processing are particles below the detection level of the particle detection system, they are not included in the total particle detection count value. When one image particle count value is associated with a particle in a still particle image, pattern recognition and identification processing is performed on the associated particle image to determine the type of particle. This operation is performed on all the images to determine the particle existence ratio of each type in the sample.

In order to determine the particle abundance ratio, the treatment is carried out only for the particles detected by the particle detection system, that is, for the corresponding particles. Particles below the particle detection level present in the captured still particle image cannot be processed because they are not included in the total number of particles detected by the particle detection system.

When the particle existence ratio is determined, the total number of particles detected is separately measured by a particle detection counter, and the total number of particles detected is multiplied by the particle existence ratio.
Calculate the number of particles by type in the sample. The particle concentration per unit volume is calculated from the number of particles of each type.
Also, a one-dimensional image sensor may be used in the particle detection system.

Next, a method of associating the number of particles in one image detected by the particle detection system with the particles existing in the still particle image will be described. As the particle detecting means, a case will be described in which a laser beam is applied to the sample flow to detect light scattering from particles in the sample.

Since the light scattering intensity from a particle is proportional to the effective scattering cross-section of the particle, it mainly contains information about the size of the particle. Therefore, the correspondence between the particle detection signal and the particle existing in the stationary particle image is obtained. Is performed by, for example, particle area or particle diameter information. The extraction of the above-mentioned characteristic parameters from the particle image can easily obtain the particle area and diameter information by using the characteristic extraction method that is performed in the ordinary pattern recognition.

When the analysis target is a predetermined detection level of the particle detection signal or more, the particle area or diameter value of the particles existing in one static particle image is calculated from the image analysis, and the calculated value is calculated from the larger one. Sort in order. Only one image particle count detected by the particle detection system is taken out from the sorted large values, and the taken out particles are detected as the particles detected by the particle detection system, and the analysis is performed as a normal identification processing target, and the remaining particles With respect to, another frame identification processing is performed. In this way, one image particle count value of the particle detection system is associated with the particles present in the captured still particle image.

Next, the separate frame identification processing will be described. After associating one image particle number in the particle detection system with the normal identification processing target particle, the classification and identification processing is executed for the unprocessed particles remaining in the captured still particle image. These untreated particles are particles that are not detected by the particle detection system and are present in the captured still particle image. However, important particles may be contained in these untreated particles, and the information cannot be discarded.

To determine the concentration and quantity of the untreated particles in the sample, the number per unit volume, actually,
By calculating the number of particles per sample volume corresponding to all the images after processing all the images and then converting the number again for all the measured samples, the number of particles in the sample, the ratio of the number of particles, etc. are determined. Thus, it is necessary to calculate the sample volume corresponding to the image processed sample. At the end of the measurement, the results of the regular identification process and the separate frame process are combined to obtain the final result.

In the separate frame identification result, there are particles that should originally be included in the regular identification result, and it is necessary to add them to the regular identification result. The result is also reported when particles that are exceptionally important but are not included in the formal identification process are included. If the particle size is small and the signal level is the same as the noise signal to detect by particle detection, or if the number of particles is very large and the analysis image becomes enormous, increase the particle detection level and analyze the static particle image. It is better to reduce the number.

By performing the above-mentioned series of image processing on the measurement sample, the existence ratio and the number of particles for each particle type in the sample can be determined. From these values, the particle concentration per unit volume in the measurement sample, or the number of particles converted into a microscopic field is known. The number of particles converted into the microscopic field is the average number of particles present in the microscopic field with an appropriate magnification, and it is necessary to set a sample volume corresponding to the microscopic field in advance.

It is necessary to maintain the measurement accuracy when another frame identification processing is executed for the particles remaining in the still particle image after associating one image particle count value of the particle detection system with the particles for normal identification processing. An image processed sample volume is needed. However, when the number of particles detected by the particle detection system is small, it is not possible to secure a sufficient number of still particle images, and as a result, the volume of the sample to be image-processed may not be sufficient.

In the separate frame identification processing, the analysis processing is performed based on the captured still particle image. Therefore, when the ratio of the original normal identification processing target particles in the sample is small, the image processing is performed. Since the particle concentration in the separate frame identification process is calculated based on the sample volume, there is a risk that errors will increase. This is because there are not so many particles in the sample, and there is no problem in analysis accuracy.

In this way, when the volume of the sample to be image-processed cannot be sufficiently obtained, in order to perform the separate frame identification, the measurement accuracy cannot be obtained sufficiently.
Or it is necessary to inform the operator that the measurement accuracy is poor and the data is unreliable even if another frame identification process is performed. However, among the particles that are subjected to separate frame identification processing,
The presence of particularly important particles must be reported regardless of measurement accuracy.

As the measurement accuracy of the separate frame identification processing, the reproducibility will be examined with reference to the total number of particles N subjected to the separate frame identification processing. Since the reproducibility in this case is 1 / SQRT (N), N = 400 when reproducibility is 5%, and N = 100 when reproducibility is 10%. However, if the reproducibility of the separate frame identification process can be set to 20% or more, N
= 25 is sufficient.

The total number of particles subjected to the separate frame identification processing is counted, and if the number of particles reaches a preset number or more, the result of the separate frame identification processing is judged to be valid. If the number of particles has not been reached, the validity of the measurement result is determined by judging it as invalid or displaying the degree of measurement accuracy.

Here, the total number of particles N to be subjected to the separate frame identification processing is N.
Although the example is shown in which the judgment is made based on the reproducibility based on the above, the reproducibility of the number of particles for each type may be used as the judgment reference, or the number of processed images may be simply used as the reference, and the sample volume of another frame identification You may think based on. These criteria are
It is necessary to set in advance before measuring.

Further, when the measurement result in the separate frame identification process does not meet the above-mentioned criteria, it is necessary to provide means for notifying the operator. It is good to add a comment or a symbol after the measurement result. Also,
The classification result is output as it is for particles that provide important information to the operator due to the presence of even one specific particle rather than the risk in the above-described separate frame identification processing. In this way, the problem that measurement results cannot be obtained for exceptional particles does not occur.

Further, when the particles to be subjected to the regular identification processing exist among the particles processed by the above-mentioned separate frame identification processing, the particle concentration is calculated after calculating the particle concentration per unit volume, and then the particles are classified into the regular identification processing particles. Improve accuracy. Even in this case, since the risk of the measurement accuracy in the above-described separate frame identification processing is taken into consideration and incorporated into the regular identification result, the incorporation does not reduce the measurement accuracy.

From the images picked up by all particle detection, the associated particles for normal identification processing are selected,
By executing the separate frame identification processing, it is possible to know the number of classifications and the classification ratio of all the particles detected.

However, among the particles that have been subjected to the regular identification processing, there are cases where particles that are not originally to be processed exist. If such an accurate correspondence cannot be obtained, the particle classification processing by pattern recognition checks whether the particles are normal classification identification processing particles that are not detected by the particle detection system. A method of putting the particles in the separate frame classification target particles or the processing is executed to the end, and at the end of the analysis, the number of particles per unit volume is calculated, and the result is subtracted from the result of the particle detection target particles.

In the above description, the analysis conditions were considered to be constant for the same analysis sample, but in the actual sample analysis, the analysis mode may be changed. Even when there are a plurality of such analysis modes, the above-described regular identification processing and separate frame identification processing are performed for each analysis mode.

For example, if a certain type of particles is analyzed so that the particles to be identified in a different frame in the first analysis mode and the particles to be normally identified in the second analysis mode, the individual analysis modes will be set after the analysis is completed. Combining the classification and identification results of
The accuracy of analysis is improved by arranging the results of particle analysis of the entire analysis sample.

[0065]

The function of each of the above technical means is as follows.
According to the configuration of the present invention, a particle sample suspended in a liquid is caused to flow in a flow cell, particles that pass through a particle detection region in the flow cell are detected, and a still image of the detected particles that have passed through the imaging region in the flow cell. The morphological classification of particles was performed by imaging the captured particle image and analyzing the captured particle image, and one image particle number count value and the particles present in the stationary particle image were made to correspond to each other. It can be confirmed whether the particles are detection particles. Therefore, even if particles smaller than the particle detection level simultaneously exist in the still particle image, they can be excluded from the detected particles, and the number of particles subjected to the entire image analysis processing and the count value of one image particle can be matched.

Further, in general, when the particle detection of the particle detection system and the principle of image processing of a still particle image are different, the image processing feature quantity is the particle size feature quantity due to light scattering from the detected particles. In addition, since there is a feature amount of image processing close to the particle detection principle of the particle detection system, it is possible to easily associate the particle detection of the particle detection system with the particles in the still particle image.

When processing is performed at a predetermined detection level or higher of the particle detection signal, characteristic numerical values such as particle area or diameter of particles existing in one static particle image are calculated by image processing, and these values are calculated. , In order from the largest, and the number of image particles above a certain signal level detected by the particle detection system from the largest sorted value, that is, by extracting only one image count value,
The detected particles detected by the particle detection system and the particles present in the stationary particle image can be associated with each other.

Individual particles can be classified and identified by image processing and pattern recognition processing with respect to particles in a static particle image associated with the detected particles detected by the particle detection system. It is possible to know the abundance ratio by particle type present in the sample.

Further, since the particle detection system is provided with a means for counting all particles larger than a certain detection signal level in the sample regardless of whether or not a still particle image is picked up, particles due to dead time generated during image capturing. It is possible to eliminate counting down of particles in image processing.

From the total number of particles in the sample and the individual particle abundance ratio which can be known from the result of image processing, the number of particles of each kind in the sample, the particle concentration per unit volume, and the number of particles converted into a unit microscope field of view. It becomes possible to know. Since it has a means for preventing false counting of particle detection signals due to flash lamp light for capturing a still particle image,
It is possible to prevent the influence of flash lamp emission.

Further, as described above, it becomes possible to perform separate frame identification processing on small particles smaller than a certain detection level in the particle detection system in the still particle image, and particles smaller than the particle detection level present in the image. Can also be classified and analyzed. Among these small particles, those that are particles for regular identification processing, those that are not particles for regular identification processing but that provide important information regarding the existence of the particles, etc. are included, and by the separate frame identification processing, It is possible to calculate the abundance ratio of processed particles, and to know the particle concentration per unit volume of the analyzed particles and the number of particles converted into a microscopic field from the sample volume corresponding to all static particle images that have been analyzed. .

In the separate frame identification processing, the measurement accuracy is judged based on the total number of processed particles of the separate frame identification particles, and the fact is reported. Therefore, the total processed particle numerical value and the preset numerical value are compared. If the measurement accuracy cannot be guaranteed by comparing the above, the fact can be reported and the error due to the separate frame identification processing can be reduced. In addition, the separate frame identification process,
Since the classification result can be output as it is, if there is more important information that even one specific particle exists, the analysis result can be obtained for the exceptional particle.

Further, in the case where the particles processed by the above-described separate frame identification processing include normal identification processing particles, the particle concentration per unit volume is calculated and then incorporated into the result of the normal identification processing. Can be improved. Also in this case, since the deterioration of the analysis accuracy in the separate frame identification processing is determined and incorporated into the result of the regular identification processing, it is possible to prevent the deterioration of the measurement accuracy due to the inclusion of the result of the separate frame identification processing.

It should be noted that these judgment criteria are set in advance before the particle analysis, and if the measurement result in the separate frame identification processing does not meet the above judgment criteria, the operator can be notified, and the analysis result can be reassured. You can judge. In the static particle image, particles that are important for sample analysis may be included, so it is not possible to discard that information, but by performing the separate frame identification process described above, these particles can be reliably detected. It can be analyzed and does not deteriorate the overall analysis accuracy.

Further, even when there are a plurality of analysis modes in which the range of particles to be analyzed is different, the normal identification processing and the separate frame identification processing can be performed for each analysis mode. For example, for a certain type of particle, if the analysis is performed so that the separate classification classification target particles can be performed in the first analysis mode and the regular classification processing can be performed in the second analysis mode, the classification identification result of each analysis mode after the analysis is completed. By combining the above, it is possible to improve the analysis accuracy by collecting the particle analysis results of the entire analysis sample.

[0076]

Embodiments of the present invention will now be described with reference to FIGS.
This will be described with reference to 4 and 5. FIG. 1 is a schematic explanatory view showing a configuration of a flow type particle image analysis apparatus used in a flow type particle image analysis method according to an embodiment of the present invention, FIG.
1 is a block diagram showing a configuration of a particle number analysis unit in the embodiment of FIG. 1, FIG. 3 is an explanatory diagram showing a relationship between a particle detection position and a stationary particle image capturing position in the embodiment of FIG.
FIG. 4 is a diagram showing a signal timing relationship for counting one image particle in the embodiment of FIG. 1, and FIG. 5 is a flow chart for explaining the processing contents of the particle number analysis unit of FIG.

In FIG. 1, 100 is a flow cell and 10 is a flow cell.
1 is an image pickup means, 102 is a particle detection means, 103 is a particle analysis means, 1 is a flash lamp, 1a is a flash lamp drive circuit, 2 is a field lens, 3 is a microscope condenser lens, 5 is a microscope objective lens, and 6 is a connection lens. Image position, 7 is a projection lens, 8 is a TV camera, 11 is a field stop, 12 is an aperture stop, 14 is a microreflector, 15 is a semiconductor laser, 17 is a collimator lens, 18 is a cylindrical lens, 19 is a reflector, 20 Is a beam splitter, 2
1 is a diaphragm, 22 is a light detection circuit, 23 is a flash lamp lighting control circuit, 24 is an AD converter, 25 is an image memory,
26 is an image processing control circuit, 27 is a feature extraction circuit, 28 is an identification circuit, 29 is a central control unit, 40 is a particle number analysis unit, 5
Reference numeral 0 is a display unit.

In FIG. 1, a flow-type particle image analysis apparatus used in the flow-type particle image analysis method according to the present embodiment includes a flow cell 100 to which a sample solution in which particles are suspended is supplied, and a flow cell 100 in the flow cell 100. The image capturing means 101 for capturing an image of the sample liquid, the particle detecting means 102 for detecting particles from the captured image, and the particle analyzing means 103 for analyzing the detected particles are provided.

The flow cell 100 comprises the sample liquid 11
0, the sheath liquid 111 is supplied, and the sample liquid 110
Form a flow surrounded by the sheath liquid 111. Then, the sample liquid flow 110 becomes a stable steady flow having a flat cross section in a direction perpendicular to the microscope optical axis 9 (described later) of the image capturing means 101 (described later), that is, a so-called sheath flow, and the center of the flow cell 100 is It flows down from the top of the paper. The flow velocity of the sample liquid flow 110 is controlled by the central control unit 29 (described later).

The image pickup means 101 has a function as a microscope and is a flash lamp 1 which is a pulse light source.
A flash lamp driving circuit 1a for causing the flash lamp 1 to emit light; a field lens 2 for collimating the pulse light flux from the flash lamp 1; and a parallel pulse light flux 10 from the field lens 2 for the sample liquid flow 110 in the flow cell 100. Microscope condenser lens 3 for focusing on the sample and the sample liquid flow 110 in the flow cell 100.
The microscope objective lens 5 for converging the pulsed light flux irradiated to the image forming an image at the image forming position 6 and the particle image at the image forming position 6 projected through the projection lens 7 are taken in by the so-called interlace method and are electric signals. A TV camera 8 for converting into an image data signal, a field diaphragm 11 and an aperture diaphragm 12 for limiting the width of the pulsed light flux 10 are provided. Above T
As the V camera 8, a CCD camera or the like having a small afterimage is generally used.

The particle detecting means 102 is a semiconductor laser 15 which is a detection light source for emitting laser light as detection light.
A collimator lens 16 for converting the laser light from the semiconductor laser 15 into a parallel laser light flux 14, a cylindrical lens 17 for focusing only one direction of the laser light flux 14 from the collimator lens 16, and a cylindrical lens 1
7 is located between the reflecting mirror 18 for reflecting the laser beam 14 from the microscope condenser lens 3 and the flow cell 100, and the laser beam 14 from the reflecting mirror 18 is located on the upstream side of the image capturing area on the sample liquid flow 110. It is composed of a micro-reflecting mirror 19 that leads to the vicinity.

Further, the laser light flux 1 by the particles is
The microscope objective lens 5 that collects the scattered light of 4 (this microscope objective lens 5 is also used as the microscope objective lens 5 for image formation of the image pickup means 101) and the microscope objective lens 5 that collects the scattered light. Beam splitter 2 that reflects scattered light
0 and the scattered light from the beam splitter 20 is limited to 2
A light detection circuit 22 which receives light via 1 and outputs an electric signal based on the intensity thereof; and a flash lamp emission control circuit 23 which operates the flash lamp drive circuit 1a based on the electric signal from the light detection circuit 22. It is equipped with.

The particle analyzing means 103 stores an AD converter 24 for converting the image data signal transferred from the TV camera 8 into a digital signal, and data based on the signal from the AD converter 24 at a predetermined address. An image memory 25, an image processing control circuit 26 that controls writing and reading of data in the image memory, and a feature extraction circuit 27 that performs image processing based on a signal from the image memory 25 to perform particle number and classification. , Identification circuit 28
The central control unit 29, the particle number analysis unit 40 for determining the number of particles in the sample liquid 110, and the display unit 50 for displaying the number of particles and the classification as a result of the above image processing.

The central control unit 29 controls the TV camera 8
Imaging conditions, sample liquid flow conditions of the flow cell 100, control of the image processing control circuit 26, storage of image processing results from the identification circuit 28, and the particle number analysis unit 40.
It is configured to exchange data with and to display on the display unit 50.

Next, the structure of the particle number analysis unit 40 will be described with reference to FIG. In the figure, the same symbols as those in FIG. 1 are the same parts, and thus detailed description thereof will be omitted. Figure 2
, 41 is a total particle counting circuit, 42 is a 1-image particle counting circuit, 43 is a 1-image particle detection timing generation circuit, 5
0 is a display unit, 61 is a one-image corresponding feature parameter data unit, 62 is a particle classification result unit, 63 is a total image number counting unit, and 6 is
4 is an analysis condition setting unit, 65 is a parameter data size rearranging unit, 66 is a corresponding unit, 67 is a particle classification result corresponding unit, 6
Reference numeral 8 is a regular classification result portion, 69 is a separate frame classification result portion, and 70 is a classification result totaling portion.

In FIG. 2, the total particle counting circuit 41 constituting the particle number analyzing section 40 receives the particle detection signal from the photodetection circuit 22 and counts the total number of detected particles during the entire analysis time. Has become. Also, one image particle counting circuit 4
2 counts particles above the particle detection level present in the image taken by flash lamp emission after particle detection. The one-image particle detection timing circuit 43 is adapted to generate a particle count time timing for one-image particle detection by the control signals from the particle detection signal 22 and the image processing control circuit 26.

Furthermore, the analysis condition setting unit 64 is set to sample analysis conditions, the total image counting unit 63 is set to count all images, that is, the count amount is incremented by 1 each time one image is processed. The result of particle classification of the one image processed particle image is set in the one-image corresponding feature parameter data unit 61 by the central control unit 29, respectively, of the parameter data associated with the detected particles in the feature amount of each classified particle. It is like this.

Further, in the parameter data size rearranging section 65, the characteristic parameters set in the one-image corresponding characteristic parameter data section 61 are arranged in order of large data value. In the data correspondence unit 66, among the parameter data arranged in descending order, from the largest parameter data, only the number set in one image particle counting circuit is determined to be the detection particles, and the rest are other than the detection particles. .

Based on the result of the data correspondence unit 66,
The particle classification result set in the particle classification result unit 62 is sorted by the particle classification result correspondence unit 67 into the target particles of the regular classification process and the target particles of the separate frame classification process, and the normal classification result unit 68 and the separate frame classification result, respectively. Set it on the part 69.

Finally, the total particle circuit 4 for counting the total number of particles
1. Based on the particle identification processing results of the total image number counting unit 63 that counts the total number of processed images, the analysis condition setting unit 64, the regular classification result unit 68, and the separate frame classification result unit 69, the classification result totaling unit 70 finally determines Processing results can be summarized. The result is transmitted to the central control unit 29,
If necessary, it is output and displayed on the display unit 50.

The operation of the flow type particle image analysis method and the flow type particle image analysis apparatus used therefor will be described. First, the image pickup means 101 and the particle detection means 10
The operation with 2 will be described. The semiconductor laser 15 constantly oscillates continuously and always observes that particles in the sample 110 pass through the detection region. The laser light flux 14 from the semiconductor laser 15 is converted into a parallel laser light flux by the collimator lens 16 and is focused by the cylindrical lens 17 in only one direction of the light flux 14.

The laser beam 14 is reflected by the reflecting mirror 18 and the minute reflecting mirror 19 and is irradiated onto the sample liquid flow 110 in the flow cell 100. This irradiation position is a particle detection position where the laser beam 14 is focused by the cylindrical lens 17, and is a position near the upstream side of the image capturing area on the sample liquid flow 110. When the particle to be analyzed crosses the laser beam 14, the laser beam 14 is scattered.

The scattered light of the laser beam is reflected by the beam splitter 20, received by the photodetector circuit 22, and converted into an electric signal based on its intensity. further,
The photodetection circuit 22 determines whether or not an electric signal for particle detection has a predetermined signal level or higher. If the detected electrical signal is equal to or higher than a predetermined signal level, it is considered that the particles to be image-processed have passed, and the detected signal is transmitted to the flash lamp emission control circuit 23 and the particle number analysis unit 40.

In the flash lamp emission control circuit 23 which receives the detection signal, the image processing target particles are
When the flash lamp 1 emits light and an image is captured when a predetermined position in the image capturing area of the V camera 8 is reached,
Based on the timing of the particle detection signal delayed by a predetermined delay time and the timing of the filled signal of the TV camera 8, the flash lamp emission timing is determined, a flash emission signal is generated and sent to the drive circuit 1a.

The predetermined delay time is determined by the distance between the particle detection position and the image capturing area, the flow velocity of the sample liquid 110, and the like. This delay time is extremely short because the distance between the particle detection position and the image capturing area is very short, and the accuracy of particle detection and analysis is not affected by the flow velocity of the sample solution and the particle concentration.

When the flash light emission signal is transmitted to the flash lamp drive circuit 1a, the flash lamp drive circuit 1a causes the flash lamp 1 to emit light. The pulsed light emitted from the flash lamp 1 travels on the optical axis 9 of the microscope, becomes parallel light by the field lens 2, is focused by the microscope condenser lens 3, and is irradiated onto the sample liquid flow 110 in the flow cell 100. The width of the pulsed light flux 10 is limited by the field stop 11 and the aperture stop 12.

The pulsed light flux applied to the sample liquid flow 110 in the flow cell 100 is condensed by the microscope objective lens 5 and forms an image at the image forming position 6. The image at the image forming position 6 is projected onto the image pickup surface of the TV camera 8 by the projection lens 7 and converted into an image data signal by the interlace method. In this way, the sample liquid 1
The static particle image of the particles in 10 is captured.
The image pickup conditions in the TV camera 8 are as follows:
Is set in advance, and the TV is
The imaging operation of the camera 8 is controlled.

The particle analysis means 103 will be described. The image data signal picked up by the TV camera 8 and converted by the interlace method is converted into a digital signal by the AD converter 24, and the data based on this is controlled by the image processing control circuit 26 and predetermined in the image memory 25. Stored at the address of.

The data stored in the image memory 25 is read out under the control of the image processing control circuit 26, input to the feature extraction circuit 27 and the identification circuit 28 to be subjected to image processing, and the central control unit 29 receives the data. The result is stored.
The contents stored in the central control unit 29 are particle identification characteristic parameters used for particle classification and particle classification result data.

The classification and identification processing of particles is automatically performed by the pattern recognition processing which is usually performed. This image processing result, measurement conditions, and information on the number of images that have been image processed,
It is sent from the central control unit 29 to the particle number analysis unit 40.
In the particle number analysis unit 40, the central control unit 2
9. Based on the particle detection signal from the photodetection circuit 22 and the control signal from the image processing control circuit 26, the correspondence between the detected particles and the particle classification result is examined, and the final classification and identification result of the particle image is obtained. Summary is done.

Next, the particle number analysis unit 40 will be described with reference to FIG. The total particle counting circuit 41 counts the particle detection signal from the photodetection circuit 22 for a signal having a level equal to or higher than a certain detection level that has passed through the particle detection system during the entire measurement time, and obtains the total number of detected particles at the end of the measurement.

On the other hand, the one-image particle counting circuit 42 counts the particles larger than the detection level present in one image picked up by the light emission of the flash lamp 1 after the particle detection.
The time timing signal for particle counting for detecting one image particle at that time is transmitted from the one image particle detection timing generation circuit 43 based on the control signal from the image processing control circuit 26.

Now, with reference to FIG. 3, the relationship between the particle detection position and the stationary particle image capture position in one image particle counting will be described. As shown in FIG. 3, Tg is the time required for particles to flow from the B end to the A end of the image, and Te is
The time required for the particles to flow from the flash lamp emission position P to the edge A of the static particle image capturing field of view, To is the time required for the particles to flow from the edge B of the image to the flash lamp emission position P, Tg = Te + To

If the particles in the measurement sample have passed the particle detection position and the particle detection signal is above a predetermined level, the detection pulse signal is output when the particles are detected. After outputting the detection pulse signal, the flash lamp 1 is caused to emit light after a lapse of a delay time Td until the particles reach a predetermined position in the static particle image capturing area.

The light emitting conditions of the flash lamp 1 are as follows. After the still particle image capture by the previous particle detection is completed, that is, the flash lamp 1 emission after the flash lamp 1 emission is prohibited, and the image signal is transferred by one field of the static particle image by two filled read signals. , The flash lamp 1 is turned on again. It is the lamp emission ready signal that makes this state,
It is made by the flash lamp control circuit 23.

The signal level of the lamp emission ready signal is "
If it is "on", the flash lamp 1 is in a light emission ready state, and if it is "off", it is in a light emission prohibition state. In FIG. 3, after y particles are detected, the flash lamp 1 is made to emit light after a lapse of Td time. However, the particles that have passed before the y-particle detection may be present in the downstream side of the flow.

As the particles, those existing at the position of the time interval Te before passage of the y particles correspond to, for example, x particles. During the time Te during which the sample particles flow from point P to end A, the lamp emission ready signal is still "o".
This is because it may not be in the n "state. This appears remarkably when the concentration of particles in the sample is high. The section where such a state occurs is the hatched section of the lamp emission ready signal in FIG. Indicated.

If the time from when the lamp light emission ready signal is turned "on" to when the lamp light is emitted becomes longer than Te,
Downstream particles, such as x particles, flow away from the image,
The shaded area disappears. One image particle counting time interval is the time T during which the particles pass through the stationary particle image capturing area.
between g.

The timing of each signal is shown with reference to FIG. In the particle detection signal, detection signals are generated from three particles x, y, and z. In order to cause the flash lamp 1 to emit light in a timely manner, a particle detection delay Td signal delayed by time Td may be transmitted after the particles are detected. Here, the lamp emission ready signal is “on” or “off”
Whether or not the lamp should emit light depends on the condition.

In FIG. 4, since the still particle image was being captured in the previous frame, the lamp light emission ready signal was not "on" from the beginning, and the lamp light emission ready signal was ready by the vertical synchronizing signal. turned on ”.
Therefore, the flash lamp does not emit light with the particle x, but emits with the next y particle, and the lamp emission ready signal is "of
f ".

As described above, the counting of the number of particles in one image cannot be started unless the flash lamp 1 emits light. However, at that time the x-particles are downstream of the flow and are in the particle image capture area. In order to count the number of image particles, Te
It is necessary to start counting from time before.

For that purpose, the particle detection delay Td signal is further delayed by Te time, and the particle detection delay Td + Te
Output a signal. The one image particle counter may count between the To + Te signals from the flash lamp emission to the time Tg. Tg is the time the particles flow between the edges of the image. After the elapse of the Tg time, one image particle count data is transferred to a necessary place.

Again, referring to FIG. 3, a further explanation will be given.
By counting the particles that have passed the time To after the flash lamp emits light at the point P, x particles are not counted and the correct number of particles present in one image cannot be known. One image particle count must be run for Tg time.

However, after the particles are detected and the flash lamp 1 emits light, the particles from the edge A of the image plane to the flash lamp emission point P are not counted and flow too much. Therefore, a particle detection delay Td + Te signal that is delayed by Te time from the particle detection delay Td signal is generated, and the one image particle counting time is set to Tg = Te + To, so that the correct one image particle counting operation is performed. The above is only one example of timing.

Again, referring to FIG. 2, the flow type particle image analysis will be described. The sample analysis condition data set in the analysis condition setting unit 64 is determined by the processing content of the classification result totaling unit 70. Basically, analysis sample volume, analysis time, sample volume equivalent to one static particle image,
Although data such as a sample volume corresponding to one field of view of the microscope and an analysis mode are required, there are some items that may be omitted from the above data depending on the classification processing content.

In the total image counting unit 63, the count amount is increased by 1 each time one image is processed, and the total number of images processed in the analysis time is stored at the end of the analysis. The particle classification result of the still particle image obtained by performing image processing on one image is the discrimination circuit 2
8 and the central control unit 29, and set in the particle classification result unit 62. At the same time, of the feature quantities used for classification and identification of the classified particles, the feature quantity closest to the detection principle of the particle detection system is set in the one-image corresponding feature parameter data section 61. This is necessary for associating one image particle count value by the detection system with the captured still particle image.

Next, the characteristic parameter data set in the one-image corresponding characteristic parameter data section 61 and corresponding to the detected particles of all the particles existing in the still particle image are arranged in the parameter data size rearranging section 65 in descending order. Be aligned. Of course, if the detection principle of the particle detection system can be handled according to the size relationship such as particle size, the sorting is performed according to the size relationship, but the sorting method may be used according to another principle.

In the corresponding unit 66, among the parameter data arranged in the descending order by the parameter data size rearranging unit 65, only the number set in the one image particle counting circuit 42 from the larger one is detected particles, and the detected particles are detected particles. Other than the rest, it is classified.

Based on the corresponding classification result of the corresponding section 66, the particle classification result set in the particle classification result section 62 is analyzed by the next particle classification result correspondence section 67 in the normal classification processing target particles having a certain signal level or higher. And the particles for another frame identification processing having a certain signal level or less, and the respective results are set in the normal classification result part 68 and the another frame classification result part 69.

The normal classification result section 68 and the separate frame classification result section 69 perform an operation of adding to the results classified and classified up to that point. When adding, the regular classification result part 6
In 8, a plurality of types of particles are added for each type. Similarly, the separate frame classification result section 69 is also added for each of the plurality of different frame classification particles. At the end of the analysis, a cumulative total is obtained for each type of all image-processed particles.

When the analysis of one sample is completed, the total number of particles above the particle detection level related to the analyzed sample is the total particle counting circuit 41, the total number of image processed images is the total image number counting unit 63, and the normal identification processing target particles. The cumulative total number of classifications is stored in the regular classification result portion 68, and the cumulative total number of particles of another frame identification processing target is stored in the separate frame classification result portion 69.

Based on this data, each of the above-mentioned stored data is sorted as the final result by the classification result totaling unit 70, and the result is returned to the central control unit 29 and output to the display unit 50 as necessary. Is displayed.

The order of particle analysis will be described with reference to the flow chart shown in FIG. When the analysis of the particle sample is started in step 1, the analysis conditions are first set in the analysis condition setting unit 64. The analysis conditions to be set include the division between regular classification and separate frame classification, analysis time, analysis amount, classification type, analysis mode, and the like.

In step 2, whether the series of particle detection / image processing has been completed for the entire sample,
It is judged whether or not it is completed, and if not completed, the process proceeds to step 3. In step 3, particle detection, particle image capturing, particle feature extraction, and particle classification identification are executed for one static particle image, and the process proceeds to step 4.

In step 4, the particle classification result and corresponding feature parameter data for one still particle image are
They are respectively stored in the particle classification result part 62 and the one-image corresponding feature parameter part, and the process proceeds to step 5. In step 5, the size of the corresponding feature parameter data is rearranged by the parameter size rearrangement unit 65, and the process proceeds to step 6.

In step 6, one image particle counting value is set in the one image particle counting circuit 42, and the process proceeds to step 7. In step 7, based on the set one image particle count value, the correspondence unit 66 processes the correspondence between the one image particle count value and the particles to be image-processed, and proceeds to step 8. In step 8, the normal classification result unit 68 performs addition processing of the analysis result of the normal identification processing target particles for each particle type, and the process proceeds to step 9.

In step 9, the additional frame classification result part 69 addition process is executed by the separate frame classification result part 69 for the analysis result of the particles to be subjected to the separate frame identification processing for each particle type, and the process returns to step 2. In step 2, it is repeated until the series of the particle detection / image processing is completed. When the analysis of all the samples is completed, the process proceeds to step 11.

In step 11, in the classification result totaling unit 70, the total number of particles above the particle detection level for the analysis sample by the total particle counting circuit 41 and the total number of images processed by the total image number counting unit 63 are set. Then, the process proceeds to step 12. In step 12, the cumulative number of classifications of the particles to be subjected to the regular identification processing is read in the regular classification result section 68, and the cumulative number of classifications of the particles to be subjected to the separate frame identification processing is read out in the separate frame classification result section 69, respectively, and the classification result is totaled In the unit 70, the regular identification process / the separate frame identification process is aggregated, the number of particles in the analysis sample is determined, and step 13
Proceed to.

In step 13, since it may be expected that the analysis error will increase depending on the number of particles subjected to the separate frame identification processing and the sample volume subjected to the image processing, the result of the separate frame identification processing is checked to determine the number of particles, Go to step 14.

In step 14, the abundance ratio of each particle is calculated from the number of particles in the result of the normal identification processing and the result of the separate frame identification processing. Next, with respect to the particles to be subjected to the regular identification process, the existing number of detected particles in the sample is determined by multiplying the existing ratio by the total number of detected particles. Next, regarding the result of the separate frame identification processing, first, the total volume of the image-processed sample is calculated from the total number of processed images and the sample volume equivalent to one still particle image. Then, the number of particles in the total volume of the analysis sample is determined.

Based on the results of the normal identification process and the separate frame identification process, the particles existing in duplicate in both are sorted into one particle type, and the non-analyzed particles are discarded. Based on these analysis results, the particle concentration in the sample is calculated, and the field-of-view converted particle number is calculated.
Go to step 15. In step 15, the analysis result is output and the process ends.

Whether data is valid or invalid can be determined based on the number of particles subjected to separate frame identification processing and the number of sample volume particles subjected to image processing. These judgment criteria need to be set in advance before analysis, but may be carried out by setting analysis conditions at the start of analysis, for example. When a sample is analyzed in a plurality of analysis modes, the operation is executed in each analysis mode, and the analysis data of each mode is integrated at the final stage to obtain the final result.

It is necessary to have a means for preventing erroneous counting of particle detection signals due to flash lamp light for capturing a still particle image. For this purpose, there is a means for eliminating the influence of flash lamp emission. As a method, an optical system is considered so that the flash lamp light does not reach the particle detector, or the flash lamp signal is not counted electrically. On the contrary, it is possible to correct the erroneous count by a method of detecting the flash lamp light and subtracting the number of times of lamp light emission at the end of the analysis. This is performed on the count result of one image particle number and total particle number.

A flow-type particle image analysis method and a flow-type particle image analysis apparatus according to an embodiment of the present invention include a biological sample or cells suspended in a liquid, blood cell components such as red blood cells and white blood cells in blood, or urine. It is effective in classifying and analyzing the urinary sediment components present in it. In particular, when counting the number of particles of urinary sediment components in urine and classifying particles, the number of particles existing in each sample may differ by several digits or more, so image processing should be performed on the particles detected by the particle detection means. Is
It is effective in knowing the information on the number of particles in the sample solution.

However, in the urinary sediment component, the types and sizes of particles are extremely diverse, and it is impossible to accurately correspond the detected particles of the particle detection system and the particles in the still particle image. Accurate particle counting, particle classification,
The particle concentration can be analyzed. In the above example, the case where a still image of the particles contained in the sample liquid continuously flowing through the flow cell is captured and the image analysis is performed is described. It can also be applied to particle image analysis of existing slide specimens.

Further, as the particle detection of the particle detection system, the case where the laser light flux from the semiconductor laser is used as the detection light and the laser light flux scattered by the particles is used has been described, but the present invention is not limited to this. It is possible to use transmitted light, a method of detecting particles by a one-dimensional image sensor, or a method of detecting particles by a change in electric resistance due to passage of particles.

Further, when the time for the sample particles to pass through the particle detection system, that is, the pulse width of the detection pulse signal converted into an electric signal is used for particle detection, the size information in the particle flow direction should be reflected. You can
However, when the particle detection of the particle detection system is associated with the captured still particle image, the particle diameter information or the projection information is appropriate. Regarding the correspondence between the particle detection and the still particle image, if the time relationship of particle detection is recorded and corresponded to the positional relationship on the still particle image, one of the particle detection signals will be correctly displayed on the still particle image. It is possible to accurately determine which particle of the particle corresponds.

In the above description, the correspondence between the number of detected particles and one still particle image is carried out for each image, but the data necessary for the number of particles corresponding to one still particle image are taken. , It is okay to sort out the correspondence after the completion. Although the interlaced scanning has been described as the TV camera imaging method in the image capturing so far, the function and effect of the invention are the same even in the non-interlaced scanning method.

[0139]

As described in detail above, according to the present invention, a flow having a particle detection system and an image pickup system respectively, and the principle of the particle detection system and the principle of particle image discrimination of the image pickup system are different. In the method for analyzing particle images and the method for analyzing particle images by flow, the detected particles and the particles in the stationary particle image are made to correspond to each other, and particles other than the particles detected by the particle detection system are present in the stationary particle image of the image capturing system. Even if the number of particles of each type in the sample, the particle abundance ratio, the particle density, the information of the concentration can be correctly known, a flow-type particle image analysis method and a flow-type particle image analysis device with good analysis accuracy are provided. be able to.

[Brief description of drawings]

FIG. 1 is a schematic explanatory view showing the configuration of a flow type particle image analysis apparatus used in a flow type particle image analysis method according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a particle number analysis unit in the embodiment of FIG.

3 is an explanatory diagram showing a relationship between a particle detection position and a stationary particle image capturing position in the embodiment of FIG.

4 is a diagram showing a signal timing relationship for counting one image particle in the embodiment of FIG. 1. FIG.

FIG. 5 is a flowchart for explaining the processing contents of the particle number analysis unit of FIG.

FIG. 6 is a schematic explanatory view of particle detection and still particle image capturing in the related art.

[Explanation of symbols]

1 Flash Lamp 1a Flash Lamp Drive Circuit 2 Field Lens 3 Microscope Condenser Lens 5 Microscope Objective Lens 6 Imaging Position 7 Projection Lens 8 TV Camera 11 Field Stop 12 Aperture Stop 14 Micro Reflector 15 Semiconductor Laser 17 Collimator Lens 18 Cylindrical Lens 19 Reflection Mirror 20 Beam splitter 21 Aperture 22 Photodetector circuit 23 Flash lamp lighting control circuit 24 AD converter 25 Image memory 26 Image processing control circuit 27 Feature extraction circuit 28 Identification circuit 29 Central control unit 40 Particle number analysis unit 41 Total particle counting circuit 42 1 image particle counting circuit 43 1 image particle detection timing generation circuit 50 display unit 61 1 image corresponding feature parameter data unit 62 particle classification result unit 63 total image number counting unit 64 analysis condition setting unit 65 parameter data Sorting unit 66 Corresponding unit 67 Particle classification result corresponding unit 68 Regular classification result unit 69 Separate frame classification result unit 70 Classification result totaling unit 100 Flow cell 101 Image capturing unit 102 Particle analyzing unit 103 Particle detecting unit 110 Sample flow 111 Sheath liquid Td Particle detection delay time Te Time required for particle image to flow from flash lamp emission position P to end A Tg Time required for particle image to flow from end B to end A To particle image to flash lamp from end B of image Time required to flow to the light emitting position P

Claims (25)

[Claims] 1. A detection step of flowing a liquid sample to be suspended in a flow cell and counting the number of particles in the sample by a particle detection system, and a static image of particles passing through an imaging region in the flow cell. The step of imaging with
In a flow type particle image analysis method comprising an analysis step of performing morphological classification of particles in the image by image analysis means, the analysis step includes one image particle count value by the particle detection system, and the stationary particle image. Extraction step for making each particle in correspondence with a particle image identification feature amount, either the particle distribution or the particle classification in the liquid sample from the corresponding relationship and the total particle count value by the particle detection system in the detection step. A flow-type particle image analysis method comprising: a particle analysis step. 2. The particle analysis step has a plurality of measurement modes for the same liquid sample, analyzes each measurement mode, and integrates all data after completion of analysis in all measurement modes. The flow type particle image analysis method according to claim 1. 3. The extraction step uses, as the particle image identification feature amount, any one of particle size information, color information, light scattering information, fluorescence intensity information, or information obtained by combining a plurality of these. The flow-type particle image analysis method according to claim 1, which is characterized in that. 4. The particle analysis step calculates either the number of particles or the ratio per sample volume for each of the entire image processed images, and then uses either the number of particles or the ratio of particles to determine the total sample volume per sample volume. 4. The flow type particle image analysis method according to claim 1, wherein the number of particles is calculated and based on the number of particles. 5. The particle analysis step determines either the number of particles in the measurement sample, the particle concentration per unit volume in the measurement sample, or the number of particles in terms of a microscopic field,
The flow-type particle image analysis method according to claim 1, wherein analysis is performed based on these. 6. The particle analysis step performs a separate frame analysis processing on the remaining particles in the still particle image that could not correspond to one image particle count value of the particle detection system, and at the end of the analysis, a particle corresponding to one image particle count value. 6. The flow-type particle image analysis method according to claim 1, wherein the separate frame processing amount is added to the analysis result of 1. 7. The flow-type particle image analysis according to claim 6, wherein the particle analysis step defines the processing when the number of processed images is small and there is not a sufficient volume to be processed to perform another frame processing. Method. 8. The particle analysis step, when the number of images to be processed is small and there is not enough processing target volume to perform another frame processing, the evaluation procedure is input and set in advance.
Or any of the flow-type particle image analysis methods described in 7 above. 9. The particle analyzing step does not perform the separate frame processing when the number of processed images is small and there is not enough processing target volume to perform the separate frame processing. Flow type particle image analysis method. 10. The particle analysis step comprises: when the number of processed images is small and the volume to be processed is small enough to perform another frame processing,
9. The flow type particle image analysis method according to claim 6, wherein another frame processing is performed, and the fact that there is a problem in the processing is output. 11. The detection step corrects an erroneous count due to pulsed light source emission for capturing a still particle image with respect to the count value of one image particle count and the total particle count by the particle detection system. Item 11. A flow type particle image analysis method according to any one of items 1 to 10. 12. A particle detection system for flowing a suspended liquid sample into a flow cell to count particles in the sample, and a particle imaging system for capturing a static image of particles passing through an imaging region in the flow cell. In a flow type particle image analysis device comprising an analysis means for performing morphological classification of particles of the stationary particle image, the analysis means is one image particle count value of the particle detection system, and each in the stationary particle image. An extraction unit that associates particles with a particle image identification feature amount, and a particle analysis unit that performs either particle distribution or particle classification in the sample from the correspondence relationship and the total number of particles counted by the particle detection system. A flow-type particle image analysis device characterized in that 13. The particle analysis unit has a plurality of measurement modes for the same measurement sample, is provided with means for analyzing each measurement mode, and integrating all data after completion of analysis in all measurement modes. The flow-type particle image analysis method according to claim 12, which is characterized in that. 14. The extraction unit uses one of particle size information, light scattering information, color information, and fluorescence intensity information, or information obtained by combining a plurality of these as the particle image identification feature amount. 15. The flow-type particle image analysis device according to claim 13, wherein the flow-type particle image analysis device is configured to perform. 15. The particle analysis unit is configured to calculate the number of particles per sample volume or particle ratio corresponding to all the image-processed images, and further to determine the number of particles per total sample volume. Claim 12
15. A flow type particle image analyzer according to any one of items 1 to 14. 16. The particle analysis unit is configured to calculate the number of particles in the measurement sample, the particle concentration per unit volume in the measurement sample, or the number of particles converted into a microscopic field. 15. The flow type particle image analysis device according to any one of items 1 to 15. 17. The particle analysis unit analyzes the remaining particles in the still particle image that could not correspond to one image particle count value of the particle detection system as a separate frame, and at the end of the analysis, a particle corresponding to one image particle count value. 17. The flow-type particle image analysis device according to claim 12, wherein the analysis result is added with the additional frame processing amount. 18. The particle analysis unit is configured to perform another frame processing when the number of processed images is small and there is not enough processing target volume to perform another frame processing.
The flow-type particle image analyzer described. 19. The particle analysis unit is configured to input and set the evaluation procedure in advance when the number of processed images is small and there is not a sufficient volume to be processed for another frame processing. Alternatively, the flow type particle image analysis device according to any one of 28. 20. The flow according to claim 17, wherein the particle analysis unit does not perform another frame processing when the number of processed images is small and there is not enough processing target volume to perform another frame processing. Particle image analyzer. 21. The particle analysis unit is configured to display or output that there is a problem in another frame processing when the number of processed images is small and the processing target volume sufficient to perform another frame processing is small. 20. The flow type particle image analyzer according to claim 17. 22. The particle detection system is configured to correct an erroneous count due to pulsed light source emission for still image capturing with respect to one image particle count value and the total particle count value. 22. A flow type particle image analyzer according to any one of claims 21 to 21. 23. The flow type particle image analysis device according to claim 12, wherein the analysis target is a biological cell. 24. The flow type particle image analysis device according to claim 23, wherein the analysis target is a blood cell component present in blood. 25. The flow-type particle image analyzer according to claim 23, wherein the analysis target is urine, particularly a urine sediment component existing in urine.