TWI453466B - Method for reducing speckle noise and apparatus using the same - Google Patents

Description

Spot noise reduction method and device using the same

The present invention relates to a method for reducing spot noise in an image and a device using the same, and more particularly to a method for reducing spot noise in an image formed using a coherent light source and an apparatus using the same.

In the coherent imaging technique, for example, laser light, ultrasonic wave, and X-ray equivalent dimming are used as the light source for forming an image. Since the coherent light source forms an image, a spot is inevitably generated in the formed image due to the roughness of the surface of the object. The spot is caused by interference such as constructive interference or destructive interference. Because the spot is not conducive to the quality of the formed image, it is considered as noise, also known as spot noise. Therefore, eliminating or reducing spot noise in images is an important issue in coherent imaging technology.

In general, a coherent imaging system basically has the configuration shown in FIG. In the configuration shown, the light emitted by the coherent light source 102 illuminates the object to be imaged, then passes through the lens 104 to form an image in the intermediate imaging plane in which the diffuser 106 is located, and then passes through the secondary lens 108 and ultimately The imaging plane forms the final image. In order to reduce the spot noise in the final imaging, a mechanically moving diffusion sheet is used as the moving diffuser 106, and the spot noise can be reduced by vibrating longitudinally or laterally, or by rotating the diffusion sheet. However, the technique of using a mechanically-moving diffuser has disadvantages such as complicated structure, poor noise reduction efficiency, difficulty in precise control, large power consumption, large variation in quality, and difficulty in reducing the overall size.

Therefore, it has been proposed to use an electronically controlled phase spatial light modulator as a variable diffuser instead of a mechanical mobile diffuser. The phase space optical modulator modulates the wavefront difference of light passing through it according to the previously determined control parameters to sequentially generate a sequence of spot images in the final imaging plane, and the spot image of the sequence is superimposed to form a spot. The noise is reduced to the final display image. When the control parameters are found, the control parameters for reducing the spot noise are found according to the criterion that the intermediate image can generate a completely independent zero-correlation spot image. Although the system using the phase optical intermodulation device can provide stable quality, low power consumption, and small size, the reduction efficiency of the spot noise is not good.

In view of the above, it is necessary to provide a coherent imaging technique in which the noise reduction efficiency of the spot noise is remarkably improved, and the quality is stable, the power consumption is low, and the size is small.

In view of the above, the present invention provides a method and a dimming imaging apparatus capable of significantly reducing spot noise.

According to an aspect of the present invention, a method for reducing a spot noise is provided for reducing spot noise of an image formed in a dimming imaging device, and a dimming imaging device includes a coherent light source for emitting a same dimming light to illuminate an object to be imaged. And a diffuser, the method comprising the steps of: generating a plurality of spot image data sequentially arranged according to a correlation coefficient for evaluating a correlation between the two spot images, such that the plurality of spot image data of the sequence are The correlation coefficient of the two-spot image data adjacent to each other is less than 0; and, according to the plurality of spot image data of the sequence, corresponding and superimposed plurality of spot images are continuously generated, thereby generating a final display image.

In addition, the step of generating a plurality of spot image data in sequence includes the following steps: generating a spot image database containing a plurality of spot image data; and calculating a basic basis according to a correlation coefficient for evaluating correlation between the two spot images Correlation coefficient of the paired spot image data formed by the spot image data and any spot image data in the spot image database, and the spot contrast value of the total image data of the paired spot image data, so that the correlation coefficient is less than 0 And the paired spot image data with the smallest spot contrast value of the total image data, and the nth spot image data of the plurality of spot image data of the combined image data, n is an integer, wherein when no sequence is generated In any spot image data, the basic spot image data is selected from any spot image data in the spot image database. Otherwise, the basic spot image data is the (n-1) spot in the plurality of spot image data of the sequence. video material.

The spot noise reduction method according to the present invention can be applied to a dimming imaging device to reduce spot noise of an image formed therefrom, the coherent light imaging comprising: a homology light source for emitting the same dimming light to illuminate an object to be imaged; An imaging optical unit that forms a first image of the object according to the same dimming of the illuminated object; and a diffuser for modulating the first image.

Preferably, the coherent light source is one selected from the group consisting of a laser source, an X source, and an ultrasonic source, and the diffuser may be a spatial light modulator.

According to the present invention, the generation of the accumulated spot image based on the negative correlation criterion can significantly reduce the spot noise in the image and obtain a highly clear image.

Embodiments of the present invention will be described in detail with reference to the drawings. However, the present invention is not limited to the following description, and the mode and details can be easily changed by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the present invention should not be construed as being limited to the description of the embodiments described below.

First, an embodiment of reducing a spot in an image in accordance with the method of the present invention will be described with reference to FIGS. 1 and 2.

1 is an example of a coherent imaging system in which a coherent imaging system 100 includes a coherent light source 102, a first imaging lens 104, a diffuser 106, and a second imaging lens 108. The coherent light source 102 emits the same dimming light to illuminate the object 103 to be imaged, and then forms an intermediate image in the intermediate imaging plane where the diffuser 106 is disposed through the first imaging lens 104, and the intermediate image performs the spot noise reduction processing via the diffuser 106 ( Details will be made later, and then a corresponding image is formed on the final imaging plane 110 through the secondary lens 108.

The coherent light source 102 can be any light source that produces the same dimming light, for example, a laser source, an ultrasonic source, an X source, and the like. In the present specification, a laser light source will be described as an example, but the present invention is not limited thereto.

In accordance with an embodiment of the present invention, the diffuser 106 is a phased spatial light modulator device, which may be, for example, a liquid crystal device panel, a liquid crystal on-chip (LCoS) device, or the like. The phase spatial light modulator device includes a plurality of pixels, and by changing the electrical signals such as voltages applied to the respective pixels, the phase of the light passing through the spatial light modulator device can be controlled to reduce the spot noise in the final image.

2(a) shows an original object 200 exemplified to form an image, and FIG. 2(b) shows a first spot image (also referred to as a first spot pattern) corresponding to the original object generated according to the method of the present invention, 2(c) shows a second spot pattern (also referred to as a second spot pattern) produced in accordance with the method of the present invention, and FIG. 2(d) shows a second spot image produced by the method of the present invention completely superimposed (or The sum image is formed on the first spot image to reduce the image of the spot noise, which is a display image of the original object visually perceived by the viewer. Here, for the sake of brevity, the total combined spot image of the two spot images in FIG. 2(d) is used as the display image. However, in reality, the display image is not limited to the combined image of the two spot images. It can be the sum of two or more spot images.

Next, a case where an image is formed using the coherent imaging system 100 shown in FIG. 1 will be described. First, light from the coherent light source 102 illuminates, for example, the object 103 shown in FIG. 2(a), and is imaged on the intermediate imaging plane via the first imaging lens 104. At this time, the diffuser 106 provided at the intermediate imaging plane performs a spot reduction process to generate a wavefront phase difference by the light of each pixel provided therein, and then forms a final image plane at the final imaging plane via the second imaging lens 108. The first spot image shown in Fig. 2(b) above. Then, within a predetermined time interval, similarly, the diffuser 106 again performs the spot noise reduction process to generate a wavefront phase difference again by the light of each pixel disposed therein, and then, via the second imaging lens. 108 forms a second spot image as shown in FIG. 2(c) above in the final imaging plane. The second spot image is superimposed on the first spot image to form the last displayed image of the original object 100. Since the predetermined time interval is selected to be smaller than the visual persistence time of the human eye or the exposure time of the image capturing device, the display image of the object that the user substantially views or captures is the second spot image. A total image formed by superimposing on the first spot image. Therefore, the degree of reduction of the spot noise in the aggregate image depends on the first spot image and the second spot image. More specifically, the degree of spot noise reduction of the combined image depends on the first spot image and the second spot image. Depends on relevance.

Regarding the correlation between the first spot image and the second spot image, in the prior art, the first spot image and the second spot image with zero correlation are used as criteria to obtain a desired display result, but the reduction is performed. The effect of the spot noise in the image is not good.

In the following, the correlation selection criteria of the two-spot image according to the present invention will be explained.

[Correlation between spot images]

Since the generation of the spot noise in the image is caused by the interference, the intensity fluctuation in the spot image (or the spot pattern) is evaluated by the spot contrast value C from the viewpoint of the intensity of the light or the amplitude of the light wave. . The spot contrast value C is defined as an average value of the standard deviation of the intensity of the spot image with respect to the intensity of the spot image, as shown by the following formula (1):

among them, And σ are the average and standard deviation of the spot image intensity, respectively.

According to the spot contrast value C, the spot image or pattern having the smaller spot contrast value indicates that the spot noise is reduced more. Images with a fixed intensity (ie, no spot fluctuations) have a minimum C value, such as zero.

The light intensity at each point in the spot image can be represented using an array. For example, assume that I ₁ and I ₂ represent an intensity array of two respective first and second spot images, respectively. For example, in the case of a spot image having a total of 1024 pixels, the array I ₁ = [P ₁ (1), P ₁ (2), ... P ₁ (1024)], the array I ₂ = [P ₂ (1 ), P ₂ (2), ... P ₂ (1024)], wherein P ₁ (1) represents the light intensity of the corresponding point in the first spot image detected by the first pixel in the spatial light modulator, P ₁ (2) represents the light intensity of the corresponding point in the first spot image detected by the second pixel, and so on.

The spot contrast value of any spot pattern can be expressed as follows:

Among them, the natural number of m=1, 2,...n.

Therefore, for the first spot image and the second spot image, their spot contrast values are C ₁ (=σ ₁ / ) and C ₂ (=σ ₂ / ).

The spot contrast value C for evaluating the spot noise in the spot image has been described above.

Correlation of two speckle images of the correlation coefficient ρ _1,2 may be defined, the correlation coefficient ρ _1,2 is defined as follows:

Wherein, I ₁ and I ₂ are intensity arrays of the first spot image and the second spot image, respectively. and And the average of the intensity of the first and second spot images, respectively, and σ ₁ and σ ₂ are the standard deviations of the intensity of the first and second spot images, respectively.

and They are the average of the intensity of each point in the arrays I ₁ and I ₂ , respectively. For example, =P ₁ (1)+P ₁ (2)+...+P ₁ (1024)/1024.

According to the formula (3), it is understood that the correlation coefficient ρ _{1,2 is} in the range of -1 to 1.

Assuming that the sum of the intensities I _S when the two images are accumulated is expressed as I _S =I ₁ +I ₂ , according to the formula (1), the sum of the spot contrast values C _{S of the} two images can be expressed as the formula (4).

Traditionally, only based on the average of the intensity standard deviation σ _s of the two images versus the sum of the intensities The ratio, that is, the C _S value shown by the formula (4), is used as an evaluation value of the spot noise reduction effect. Conventional techniques use a two-spot image with zero correlation (ie, ρ _1,2 =0) as a criterion for reducing spot noise. In short, the conventional technique performs the reduction of the spot noise under the premise that the two-spot image is completely independent, that is, zero correlation. However, according to the well-known zero correlation criterion, satisfactory spot noise reduction effects cannot be obtained.

It is not possible to meet the requirements of reduced speckle noise by relying only on zero correlation and not considering other well-known conventional theories. For this reason, the inventors have long studied, discovered and completed the principle of negative correlation, and invented the method of reducing the noise of the spot noise using the negative correlation as a criterion and the apparatus using the same, and the effect thereof is remarkably improved.

Hereinafter, the principle of the negative correlation criterion of the inventors of the present invention will be further explained.

The average intensity of the sum image of the two images Add the average values of the intensities of the two separate images as shown in the following equation (5):

Then, the variation value is obtained according to the intensity of the two images. , as shown in the following (6). Since the square of the standard deviation is the variation value, the variation value obtained from equation (6) , that is, the standard deviation σ _s is obtained.

Next, when evaluating the spot contrast value C _S of the total image, the average value of the sum image obtained according to the formula (4) and the equation (5) is obtained. And the standard deviation σ _s obtained according to the formula (6), the spot contrast value C _{S of} the total image is expressed by the following formula (7):

From equation (7), it is clear that the spot contrast value C _S of the combined image depends on the intensity average of the first and second spot images. with And the spot contrast values C ₁ and C ₂ , and the correlation coefficients ρ _{1,2 of the} first and second spot images. Therefore, in evaluating the spot noise reduction effect, according to the present invention, various possible correlations of the two-spot image can be considered, and it is possible to achieve excellent spot noise reduction efficiency.

As described above, when performing the homology imaging, the time interval for forming the two-spot image is relatively short, and the intensity of the coherent light source for forming the two-spot image remains unchanged, so that the two-spot image has an equal average intensity, that is, , . Therefore, the formula (7) can be simplified as shown in the following formula (8):

According to the present invention, the spot noise reduction effect of the sum of the two spot patterns is evaluated using the suppression factor _Cf , and the suppression factor _{Cf is} defined as follows:

According to the invention, C _avg is the average of the spot contrast values C ₁ and C ₂ of the two-spot pattern, that is, C _avg = (C ₁ + C ₂ )/2.

Therefore, the suppression factor C _f can be expressed as in the following formula (10):

According to equation (10), negative correlation, zero correlation, and positive correlation can be considered to determine the two separate spot images to be superimposed. Figure 3 shows the relationship of the suppression factor C _f plotted against equation (10) with respect to the correlation coefficient ρ _1,2 , which is based on having equal average intensities (i.e., And the conditions of the fully developed two spot images (ie, C ₁ = C ₂ =1).

Fig. 3 is a graph showing the relationship between the suppression factor C _f and ρ _1,2 according to the formula (10). It can be clearly seen from Fig. 3 that when the two spot images with fully developed spots are irrelevant (or independent), that is, when ρ _1,2 =0, their total intensity suppression factor C _f is equal to 1/ , consistent with traditional theoretical values. When ρ _1,2 =1 and C ₁ =C ₂ , then C _f =1, which means that the combined spot image is the same as each spot image in the two spot images, so no spot noise is completely suppressed. However, when ρ _1,2 =-1 and C ₁ =C ₂ , then C _f =0, the image of the spot representing the total does not contain intensity fluctuations, that is, the spot noise is completely eliminated. The case where ρ _{1,2 is} equal to zero (0) means that the two-spot image has zero correlation, the case where ρ _{1,2 is} greater than zero (0) is called positive correlation, and the case where ρ _{1,2 is} smaller than zero (0) is called negative correlation.

As is clear from Fig. 3, when the correlation is positive, the suppression factor is greater than 1/ In the case of negative correlation, the suppression factor is smaller, representing higher spot noise suppression efficiency. Therefore, the negatively correlated two spot images will provide significantly improved spot noise suppression efficiency. In addition, the smaller the suppression factor, the higher the efficiency of the spot noise suppression.

The negative correlation principle created by the inventors according to the present invention has been clearly explained above. Next, the operation processing for reducing the spot noise drop according to the negative correlation principle of the present invention will be explained with reference to the flowchart of FIG.

As shown in FIG. 4, in process 500, a negative correlation paired spot image of a predetermined number M is generated. Here, M=10 is taken as an example, but M is not particularly limited and may be any self-integer value of at least not less than 2. Depending on the design requirements, the needs and resources of the imaging equipment, etc.

FIG. 5 shows a flow diagram of process 500. In step S510, initial spot image data is randomly generated as basic spot image data. The spot image data is, for example, image intensity, phase, and the like in the form of an array. Therefore, the initial spot image data is randomly generated as described herein, and the phase value θ in the range of 0 to 2π is randomly generated to represent the spot image. . Thus, a phase element array corresponding to the initial spot image is generated in a manner that produces a random number of random numbers.

Next, in step S520, a plurality of spot image data of a predetermined number T are randomly generated as comparative spot image data. The number of comparison spot images may be determined depending on application requirements and resources, and is not particularly limited. For example, T=1,000,000.

Next, in step S530, according to equations (3) and (10), calculate a C _f value of each of the plurality of comparative spot image data when paired with the basic spot image data to form a combined image, and A paired image having the most negative correlation coefficient ρ _1,2 (the smallest value of C _f ) is selected therefrom, and the comparative spot image data is stored. Then, the selected comparative spot image data and the initial spot image data are accumulated (superimposed) to obtain the combined image data, which can be used to form a combined image. Increase the number of groups M by 1.

Next, it proceeds to step S540 to determine whether the number of groups M reaches a predetermined number, for example, 10. If the predetermined number has not been reached, the process proceeds to step S550.

In step S550, the aggregated image data acquired in step S540 is set as the basic spot image data. Then, returning to step S520, it continues to find the next set of negative correlation paired spot image data having the smallest C _f value until M reaches a predetermined number of times.

After processing 500 is completed, a predetermined set of M pairs of spot image data having a negative correlation is obtained. The spot image data (eg, phase, etc.) can be converted into a control signal that controls the diffuser correspondingly to control the diffuser to form an intermediate image of the solid object into a spot image and superimpose to form a combined image.

Table 1 shows the 10 negatively correlated spot image data obtained under Condition 1 of M = 10 and T = 1,000,000 according to the process 500, and their spot suppression factor _Cf and correlation coefficient ρ. As shown in Table 1, the two-spot image pairing represented by m=1 and m=2 has a correlation coefficient of -0.262 and C _f =0.600, and m=2 and m=3 represent the two-spot image pairing. The correlation coefficient is -0.275 and C _f = 0.448, and so on. It should be noted that, according to the iterative operation of the process 500 (steps S520 to S550), the mth image data is the total image data of the first image to the m-1th image data.

Table 1

In addition, under zero-correlation conditions, spot image search for reducing spot noise is performed to compare the results of the negative correlation implementation according to the present invention, as shown in Table 2.

Table 2

From Tables 1 and 2, it is clear that at m=4, the Cs value obtained by the embodiment of the negative correlation criterion according to the present invention is 0.334, which is significantly lower than the 0.485 obtained by the zero correlation (about 31.8% lower). . At m=10, the Cs values were 0.164 and 0.31.5, respectively, and were significantly lower (about 48% lower).

As apparent from the above, according to the negative correlation principle of the present invention, the spot noise can be remarkably reduced.

As described above, after the processing 500 acquires a plurality of negatively correlated spot image data, the process proceeds to process 600.

In the process 600, further selecting, from the plurality of negative correlation spot data obtained in the process 500, a plurality of negative correlation spot image data required to generate a corresponding plurality of negative correlation spot images and superimposing them to form a final display. Image. Preferably, the spot image data of the negative correlation coefficient is selected to be smaller (more negative) to form a corresponding spot image. The number of selected spot images depends on the resources and design of the same dimming imaging device to be applied. For example, the number of spot patterns to be selected is determined depending on the response speed of the diffuser in the dimming imaging device, so that the spot pattern superimposition can be completed within the visual retention time to form a total display of the final display seen by the user. image.

The embodiment of the flare noise reduction method according to the present invention has been described above.

When the negative correlation spot image data obtained by the spot noise reduction method according to the present invention is input to the same dimming imaging device 100 as shown in FIG. 1 as a control parameter for controlling the diffuser 106, the diffuser 106 according to the control parameter The intermediate image is processed to form a spot image 110. When the phase values of the four negatively correlated spot images are input to the same dimming imaging device 100, the final display image is a superimposed image in which the four spot images are superimposed, and the spot noise is significantly reduced.

The phase data of the desired negative correlation spot image can be obtained in accordance with the spot noise method according to the present invention in a computing system such as a computer independent of the dimming imaging device 100. These data are then input to the same dimming imaging device 100 as a control diffuser 106 to reduce spot noise in the displayed image. These data can be built into the same dimming imaging device or input from their communication port or input port, or stored in a computer readable medium, and then read by the same dimming imaging device.

According to the present invention, the use of M uncorrelated spot patterns reduces the suppression factor ( _Cf ) to 1/M ^1/2 . For example, the suppression factor using four negative correlation spot patterns (ρ ~ -0.274) is about 31.8% lower than the suppression factor of the four zero correlation spot images. The suppression factor using ten negative correlation spot patterns (ρ ~ -0.284) is about 48% lower than the suppression factor of the four zero correlation spot images. Therefore, the present invention can significantly reduce the spot noise in the image and obtain a highly clear image.

In the above, the present invention is described by way of examples, but is not intended to limit the scope of the invention. The novel embodiments disclosed herein may be embodied in various other forms and the embodiments of the embodiments described herein may be omitted, without departing from the spirit of the invention. The scope of the appended claims and their equivalents are intended to cover such forms or modifications within the scope and spirit of the invention.

100. . . Same dimming imaging device

102. . . Coherent light source

103. . . object

104. . . First imaging lens

106. . . Diffuser

108. . . Second imaging lens

110. . . Final imaging plane

1 is an example of the architecture of a dimming imaging device;

2(a) to 2(d) respectively show an original object to be imaged by the same dimming imaging device, a first spot image corresponding to the original object, a second spot image corresponding to the original object, and a combined image for respectively An embodiment of a method for reducing flare noise in an image according to the present invention is illustrated;

3 is a theoretical diagram showing a correlation coefficient between a suppression factor _Cf of a total image and a two-spot image according to the present invention;

4 is a flow chart illustrating an embodiment of a method for reducing a spot noise according to the present invention; and

Fig. 5 is a flow chart for explaining the negative correlation spot image generation processing shown in Fig. 4.

Claims (10)

A spot noise reduction method for reducing spot noise of an image formed in a same dimming imaging device, the same dimming imaging device comprising a coherent light source for emitting the same dimming light to illuminate an object to be imaged, and a diffuser, the method The method includes the following steps: generating a plurality of spot image data of the sequence according to a correlation coefficient for evaluating correlation between the two spot images, such that the two spot image data adjacent to each other in the plurality of spot image data of the sequence The correlation coefficient is less than 0; and, according to the plurality of spot image data of the sequence, corresponding and superimposed plurality of spot images are continuously generated, thereby generating a final display image. The method of claim 1, wherein the step of generating a plurality of spot image data of the sequence further comprises the steps of: generating a spot image database comprising a plurality of spot image data; Correlation coefficient of the correlation, calculating a correlation coefficient between the pair of spot image data formed by the basic spot image data and any spot image data in the spot image database, and calculating the total image data of the pair of spot image data The contrast value of the spot is such that the paired spot image data having a correlation coefficient less than 0 and the smallest is obtained, and the aggregated image data is used as the nth spot image data of the plurality of spot image data of the sequence, n is an integer, Wherein, when any spot image data of the plurality of spot image data of the sequence is not generated, the basic spot image data is selected from any spot image data in the spot image database; otherwise, the basic spot image data is The (n-1)th spot image data of the plurality of spot image data of the sequence. For example, the method of claim 2, wherein the spot image data The numerous image data contained in the library are generated independently and randomly from each other. The method of claim 2, wherein the variation value of the pair of spot image data is calculated, and the spot contrast value of the aggregate image is calculated according to the calculated variation value. The method of claim 1, wherein the final display image is generated within a human visual persistence time. The method of claim 1, wherein the coherent light source is one selected from the group consisting of a laser source, an X source, and an ultrasonic source. The method of claim 1, wherein the diffuser is a spatial light modulator. An identical dimming imaging device comprising: a coherent light source for emitting the same dimming light to illuminate an object to be imaged; a first imaging optical unit that forms a first image of the object according to the same dimming light that illuminates the object; and a diffuser for The first image is modulated, wherein the same dimming imaging device uses the method of any one of items 1 to 5 of the patent application to reduce spot noise in the formed image. The apparatus of claim 8, wherein the coherent light source is one selected from the group consisting of a laser light source, an X light source, and an ultrasonic source. The device of claim 8, wherein the diffuser is a spatial light modulator.