JP2009284270A - Image processing apparatus, imaging device, image processing method, and program - Google Patents

Abstract

Provided are an image processing apparatus, an imaging apparatus, an image processing method, and a program capable of realizing exposure control and the like that effectively use a detection result of a moving region in fixed-point shooting or quasi-fixed-point shooting using a monitoring camera or the like. To do.
At least an imaging unit that inputs an image based on control from the outside, a detection unit that detects a moving area in the screen from the image and outputs detection information, and a past output from the detection unit The storage unit 103 stores detection information, and the control unit 104 performs predetermined control on the imaging unit 101 based on past detection information stored in the storage unit 103.
[Selection] Figure 4

Description

  The present invention relates to an image processing apparatus, an imaging apparatus, an image processing method, and a program to be applied such as a surveillance camera.

  A technique is known in which a person's face is detected from an input image of a camera and a person who is a subject is clearly photographed by performing exposure control, white balance control, focus control, and the like based on the detection result (for example, Patent Documents). 1, 2, 3).

  According to the technique described in Patent Document 1, a face is detected in advance from input image data, and exposure control is performed by highly evaluating the weight of a photometric result for an area where the face is detected. The exposure of the person in the image to be photographed is made appropriate.

Further, according to the technique described in Patent Document 2, an image for face detection is held, and AF, AE, and AWB evaluation values are derived for this image.
For this reason, according to the technique described in Patent Document 2, when a person moves somewhat due to the time difference between the face detection target image and the shooting target image, more accurate control than the technique described in Patent Document 1 can be expected. .

On the other hand, according to the technique described in Patent Document 3, the accuracy of the detection result is improved by performing face detection on a plurality of images close in time with respect to subjects that can be regarded as substantially the same, and high accuracy. Realize exposure control.
JP 2003-107555 A JP 2005-318554 A JP 2007-82026 A

  What is common in the above-described techniques is that the time difference between the face detection target image and the image on which various controls are performed based on the detection result, that is, the shooting target image is relatively small, and there is no large difference between the two images. This is a premise.

In other words, all of the above-described techniques detect a face area existing in the image to be imaged itself, and perform various controls so that the area becomes clear.
In other words, these techniques can be said to be effective techniques mainly for taking still images, for example, when the subject itself is taken to be photographed and the camera is almost stationary and facing the camera.

  On the other hand, the above-described technique is difficult to apply to a case where the subject is basically not stationary, such as when a moving image is captured with a surveillance camera.

  FIG. 1A is a diagram illustrating an example of an input image at a certain time. FIG. 1B is a diagram showing an example of the input image after a slight time has elapsed from the time of the input image shown in FIG.

In the techniques described in Patent Documents 1 and 3, when a face is detected at a position indicated by reference numeral F21 in FIG. 1A, a face exists at a position F31 equivalent to the position F21 in FIG. It is assumed that it will be detected.
However, since the position where the face actually exists in FIG. 1B is the position indicated by the symbol F32, these techniques do not function effectively.

Further, at the time of acquiring the face detection target image, it is unknown where the face region is, and therefore exposure control that can detect the face is not always performed.
For example, even when photographing the same place, it is generally possible that a face is detected under clear daylight and no face is detected under a streetlight at night.

  FIG. 2A is a diagram illustrating an example of an image taken in the daytime. FIG. 2B is a diagram showing an example of an image taken at night at the same angle as FIG.

Although depending on the technique, in order to detect a human face from an image, a predetermined contrast is required in the face area.
However, in the image as shown in FIG. 2B, exposure control may be performed such that the contrast of the face area is lowered due to the influence of the bright area around the street lamp 1, and the detection may fail. is there.
If the face detection target image cannot be detected even if a face region exists in the face detection target image, none of the above-described techniques can be used.

Further, a face is not always detected in all still images constituting a series of moving images (scenes) in which a certain person is shown. That is, the current face detection technology is not 100% detection rate.
When the above-described technique is used in such a case, exposure control that is largely different between a still image in which a face is detected and a still image in which the face is not detected is performed.
As a result, in the above technique, the exposure is not stable in one scene, and flickering may occur when viewed as a moving image.

  FIGS. 3A to 3E are diagrams showing a group of frames constituting an example of a scene in which a person appears.

Here, it is assumed that the face can be detected only in the frame FLb in FIG. 3B and the frame FLd in FIG.
Further, it is assumed that the exposure control of the frame FLb in FIG. 3B is performed based on the photometry result of the frame FLa in FIG.
Similarly, it is assumed that the exposure control of the frame FLc in FIG. 3C is performed based on the photometry result of the frame FLb in FIG.
Assume that the exposure control of the frame FLd in FIG. 3D is performed based on the photometric result of the frame FLc in FIG.
Assume that the exposure control of the frame FLe in FIG. 3E is performed based on the photometric result of the frame FLd in FIG.
When a face is detected by the above-described technique, exposure control is performed so that the image becomes brighter than when a face is not detected.

Then, as shown in FIGS. 3A to 3E, the exposure is controlled so that the frames FLb and FLd are relatively dark and the frames FLc and FLe are relatively bright. An unstable situation occurs.
It is predicted that flicker will be perceived in the moving image shot in this way. However, flickering can be reduced by reducing the difference between the weight of the photometric result in the face area and the weight of the photometric result in the other areas.
However, this also impairs the effects of the technology described above.

  The present invention relates to an image processing apparatus, an imaging apparatus, an image processing method, and a program capable of realizing exposure control and the like that effectively use a detection result of a moving area in fixed-point shooting or quasi-fixed-point shooting using a surveillance camera or the like. It is to provide.

  An image processing apparatus according to a first aspect of the present invention detects a moving area in a screen from an image captured in a state according to control information, outputs a detection information, and a past output by the detection unit And a control unit that generates the control information based on the past detection information stored in the storage unit and controls the imaging state in accordance with the control information.

  An imaging device according to a second aspect of the present invention includes at least an imaging unit that captures an image based on control from the outside, and detects a moving area in the screen from the image captured by the imaging unit and outputs detection information A detection unit that accumulates past detection information output by the detection unit, and a control unit that performs predetermined control on the imaging unit based on past detection information history accumulated in the storage unit. Have.

  Preferably, the imaging unit has a function of acquiring an evaluation value related to imaging of an image, outputting the evaluation value to the control unit, and changing an imaging condition according to control information of the control unit, and the control unit includes the storage It is determined whether or not a moving region has been detected in the past from the detection information history by the unit, and if it is determined that no moving region has been detected in the past, the evaluation value is acquired from the entire screen or the center of the screen, When it is determined that a moving area has been detected in the past, the imaging unit is controlled to acquire the evaluation value for the area in which the moving area has been detected in the past, and the evaluation value obtained by the imaging unit is determined. Control information is output to the imaging unit to change imaging conditions.

  Preferably, the control unit performs control so that an evaluation value is acquired from the entire screen or the central portion of the screen until the ratio of the area in which the moving area has been detected in the past occupies the predetermined value or more.

  Preferably, the imaging unit has a function of acquiring an evaluation value related to imaging of an image, outputting the evaluation value to the control unit, and changing an imaging condition according to control information of the control unit, and the control unit includes the imaging Control to acquire a partial evaluation value for each area of the screen divided into a plurality of areas in the section, and each area from which the partial evaluation value is acquired is detected as a moving area based on the detection information history of the storage section. If it is determined whether the frequency of the region is high and it is determined that the detection frequency is high, the partial evaluation value is updated by relatively weighting the partial evaluation value, and the detection frequency is determined to be low The partial evaluation value is updated by relatively low weighting, the evaluation value of the entire screen is obtained by taking into account the updated partial evaluation value, and control information corresponding to the evaluation value of the entire screen is stored in the imaging unit. Output and change imaging conditions.

  Suitably, the said control part sets the value of the weight for updating the said partial evaluation value according to the detection frequency of a moving region.

  Preferably, an erasing unit that erases part or all of past detection information stored in the storage unit under a predetermined condition is included.

  Preferably, the predetermined condition is that a predetermined time has elapsed since the previous erasure.

  Preferably, the imaging unit is capable of imaging by changing a shooting angle, and outputs current angle information to the erasing unit, and the erasing unit outputs an erasing signal when a change in the shooting angle is detected from the angle information. The storage unit outputs to the storage unit, and the storage unit receives the erase signal and discards at least a part of the stored past detection information.

  Preferably, the image processing unit includes a switching unit that performs switching processing according to angle information, the imaging unit is capable of imaging by changing a shooting angle, outputs current angle information to the switching unit, and the storage unit Includes independent N storage units corresponding to the N types of angles (N is a natural number), and the switching unit switches to a storage unit corresponding to the angle information to store and detect corresponding detection information. The information history is input to the control unit.

Preferably, based on the detection information by the detection unit and the detection information history of the storage unit, a convergence degree acquisition unit for acquiring a convergence level of past detection information history stored in the storage unit;
A range setting unit that generates range information indicating a range of an input image that is a target of moving region detection from the convergence degree and the detection information history, and the detection unit is a range set by the range setting unit. Detect moving area.

  Preferably, when the convergence degree acquiring unit recognizes that at least one moving region has been detected by referring to detection information from the detecting unit, the moving region detected from the detection information history of the storage unit has moved in the past. It is determined whether or not the region is included in the detected region, and when it is determined that the region is not included, the convergence is decreased, and when it is determined that the region is included, the convergence is increased and the range is determined. Output to the setting section.

  Preferably, the range setting unit sets the range so that the entire screen becomes the detection target every time the detection of the moving area is performed a predetermined number of times even after the detection information history converges.

  An image processing method according to a third aspect of the present invention includes a detection step of detecting a moving area in a screen from an image captured in a state corresponding to control information, obtaining detection information, and a past obtained in the detection step. An accumulation step for accumulating the detected information, a generation step for generating the control information based on the past detection information accumulated in the accumulation step, and a control step for controlling the imaging state according to the control information .

  According to a fourth aspect of the present invention, a detection process for detecting a moving area in a screen from an image captured in a state corresponding to control information and obtaining detection information, and past detection information obtained by the detection process are obtained. An image process including an accumulation process to be accumulated, a generation process to generate the control information based on past detection information accumulated in the accumulation process, and a control process to control an imaging state according to the control information. This is a program to be executed.

According to the present invention, the detection unit detects a moving area in the screen from an image captured in a state corresponding to the control information, and outputs the detection information to the storage unit.
In the storage unit, past detection information output by the detection unit is stored.
Then, in the control unit, control information is generated based on past detection information stored in the storage unit, and the imaging state is controlled in accordance with this control information.

  ADVANTAGE OF THE INVENTION According to this invention, the exposure control etc. which utilized the detection result of a moving area effectively in fixed-point imaging | photography or semi-fixed-point imaging | photography with a surveillance camera etc. are realizable.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In the present embodiment, an imaging apparatus (camera apparatus) having an image processing apparatus that realizes exposure control that effectively uses face detection results in fixed-point shooting or quasi-fixed-point shooting with a monitoring camera or the like will be described.

  In the present embodiment, first, the background and outline of the present invention will be described, and then a specific configuration example will be described.

There is a need to clearly see a suspicious person in surveillance camera surveillance applications.
In the existing technology, a still image is mainly targeted, and control is performed so that the area where the face is present is clear now.
Surveillance cameras and the like perform fixed-point shooting or semi-fixed-point shooting that switches a preset angle manually or automatically, and the shooting time is often 24 hours a day, 365 days a year.

For example, when used outdoors, the lighting conditions of the subject change day and night. In other words, there is a possibility that a face may not be detected under night illumination conditions, and if a face is not detected, it is difficult to apply an existing technique.
In a surveillance camera, it is desirable that a region where a face can exist is clearly visible even if the face cannot be detected.

Further, depending on the condition of light and shadow, there may be a face that can be detected and a face that cannot be detected even within the same screen.
In a surveillance camera, it is desirable that all areas where people can exist are clearly visible. In the surveillance camera, even if only a specific person on the screen can be clearly seen, it is not necessary that other persons are unclear.

Generally, in surveillance applications, it is not conscious that a subject person is basically photographed. The subject may move violently, but existing technology assumes that the subject is almost stationary.
That is, it is difficult to apply the existing technology as it is for monitoring purposes.

Therefore, in the present invention, first, the screen is divided into “area where a person can exist” and “area where no person exists”.
For this purpose, in the present embodiment, face detection is performed, detection results are accumulated, and a map indicating a portion where a person can exist (= portion where a face cannot exist) is generated.
In the present embodiment, the detected content does not have to be a face as long as the position of the person can be specified. If the angle is the same, there are many cases where people can be expected to have almost the same area at daytime and nighttime. It is generated and the night exposure control is performed using the map.

In the present embodiment, the control is performed in the direction of increasing the visibility of the entire “region where a person can exist” described above.
In this case, even if there is a face that cannot be detected due to the difference in illumination conditions, the face is controlled in a direction that improves the visibility.
If the map update converges after shooting for a long time, for example, if the place where the face can exist and the place where the face does not exist is stable and the map update converges, the face detection is performed only in the place where the face can exist. However, an equivalent result to that obtained when face detection is performed on the entire screen can be obtained.

In addition, even when used in combination with existing technology that controls from the current face position, it is possible to reduce power consumption for detection.
Even if the detection speed of face detection is slow or the detection rate of face detection is low, it only takes time for map update convergence, there is no difference in the final map, and the power consumption for face detection Can be suppressed.

  Next, specific embodiments of the present invention will be described.

<First Embodiment>
First, a first embodiment of the present invention will be described. In the first embodiment, it is assumed that fixed-point shooting, that is, shooting at a constant angle is always performed.

  FIG. 4 is a block diagram illustrating an example of a configuration of an imaging apparatus (camera apparatus) to which the image processing apparatus according to the first embodiment of the present invention is applied.

As shown in FIG. 4, the imaging apparatus 100 includes an imaging unit 101, a detection unit 102, a storage unit 103, and a control unit 104.
Among these components, the detection unit 102, the storage unit 103, and the control unit 104 form an image processing apparatus.

The imaging unit 101 receives a light wave composed of reflected light from a subject, converts the light wave into an image, and outputs the image to a processing system such as the input image IM and the subsequent detection unit 102.
The imaging unit 101 obtains, for example, a control evaluation value (exposure evaluation value) EST for exposure control simultaneously with the image, and outputs the control evaluation value to the control unit 104.
The imaging unit 101 changes the exposure amount in the output image based on the control information CTL from the external control unit 104.

The detection unit 102 receives the image IM from the imaging unit 101, detects the face of a person who is a subject from the input image IM, and outputs detection information DTC to the storage unit 103. Here, the face corresponds to a moving area (structure) in the screen.
Here, the detection information is information for specifying a face region in the image, such as the position and size where the face is specifically detected.
Note that the detection speed required for the detection unit 102 may be slower than the speed required for the existing technology. The detection rate may be low, but it is desirable that the false detection rate is low.
For example, the detection unit 102 calculates an evaluation value indicating the likelihood of a face for each area. Can be applied. In this case, it is possible to reduce the false detection rate by setting the predetermined threshold value high.

The storage unit 103 receives the detection information DTC from the detection unit 102, converts it into a predetermined format, and stores it.
Hereinafter, the past detection information stored in the storage unit 103 is referred to as “detection information history”.

  The control unit 104 has a function of inputting the exposure evaluation value EST from the imaging unit 101 and the detection information history DTCH stored in the storage unit 103 and controlling the exposure amount in the imaging unit 101 based on these.

  FIG. 5 is a block diagram illustrating an example of a detailed configuration of the imaging unit 101 according to the present embodiment.

  The imaging unit 101 includes at least a lens 1011, an aperture 1012, an imaging element 1013 including a CCD or a CMOS sensor, an analog-digital (AD) converter 1014, and an exposure evaluation value calculation unit 1015.

In the imaging unit 101, the input light wave forms an image of the subject on the imaging element 1013 through the lens 1011. The image sensor 1013 converts light from the subject into an analog electrical signal, and the AD converter 1014 further converts the analog electrical signal into a digital signal, which is finally output as digital image data (IM).
In this process, the exposure evaluation value calculation unit 1015 calculates an exposure evaluation value for the output image data and outputs it to the control unit 104.

  The imaging unit 101 further includes a diaphragm 1012, an electronic shutter / mechanical shutter, a variable gain amplifier, and the like as necessary. The imaging unit 101 can appropriately adjust the exposure amount of the output image by changing the aperture of the diaphragm 1012, the shutter opening time, the gain amount, and the like based on the control information CTL from the control unit 104. .

In the imaging unit 101 shown in FIG. 5, only the aperture 1012 is provided for simplicity. Based on the control information CTL from the outside, by increasing or decreasing the aperture of the aperture 1012, the amount of light irradiated to the image sensor 1013 and thus the exposure amount can be adjusted.
Further, by another control information, the exposure evaluation value calculation unit 1015 may be controlled to acquire the exposure evaluation value EST only from a predetermined region on the image sensor 1013.
Alternatively, the exposure evaluation value calculation unit 1015 may control the partial evaluation value of each region multiplied by an independent weight for each region and then add all the partial evaluation values as the exposure evaluation value EST.

  FIG. 6 is a flowchart illustrating an example of an operation flow from inputting a light wave to constructing a detection information history in the storage unit 103.

After starting the processing, in step ST1, the imaging unit 101 converts the input light wave into image data IM and outputs the image data IM to the detection unit 102.
Next, in step ST <b> 2, the detection unit 102 detects a face area in the image data, and outputs detection information DTC to the storage unit 103.
Finally, in step S3, the detection information history DTCH in the storage unit 103 is updated with the detection information DTC by the detection unit 103, and then the process ends.

  FIGS. 7A and 7B are diagrams showing an example of a face detection situation and a detection information history format in the present embodiment.

FIG. 7A is a diagram illustrating an example of a face detection situation in a series of images.
In FIG. 7A, each of the images F01 to F06 is divided into 5 × 5 vertical and horizontal areas, the areas where the face is detected are hatched, and the other areas are not hatched. .

FIG. 7B is a diagram illustrating an example of the format of the detection information history in the storage unit 103.
Here, FIG. 7B shows a state where the detection results for the six images shown in FIG. 7A are accumulated.
In FIG. 7 (b), a set of information specifying the image in which the face is detected and the coordinates of the face area in the image is set as one entry, and the detection information for a certain past period is held in the form of a so-called queue. Has been.
In FIG. 7B, the detection information DTC includes the first face area coordinates (3, 2) of the frame F01, the second face area coordinates (1, 3) of the frame F01, and the first face of the frame F02. The area coordinates (1, 4) and the first face area coordinates (2, 4) of the frame F02 are illustrated.
Further, in FIG. 7B, the detection information DTC includes the first face area coordinates (2, 3) of the frame F05, the first face area coordinates (2, 2) of the frame F06, and the second face area of the frame F06. The face area coordinates (3, 3) are illustrated.
It is assumed that when the queue is full, the old detection information DTC is basically discarded in order.

  FIGS. 8A and 8B are diagrams showing another example of the face detection status and the detection information history format in the present embodiment.

FIG. 8A is a diagram illustrating another example of the situation of face detection in a series of images.
In FIG. 8A, as in FIG. 7A, each of the images F11 to F16 is divided into 5 × 5 vertical and horizontal areas, the area where the face is detected is hatched, and the other areas are hatched. Shall not be given.

FIG. 8B is a diagram illustrating another example of the format of the detection information history in the storage unit 103.
In the example of FIG. 8B, unlike the format shown in FIG. 7B, the past detection information is held in the form of a frequency map for face detection in each region of the image.
It should be noted that a predetermined upper limit exists for the value that the frequency can take, and the frequency exceeding the upper limit is clipped at the upper limit.
For example, if it is only necessary to distinguish whether or not a face has been detected once in a predetermined area, the values that can be taken are only 0 and 1, and the area in which a face has already been detected in the past (that is, the frequency) Even when a face is detected again in the area (1), the frequency remains at 1.
In this case, the accumulation unit 103 accumulates past detection information in the form of binary information indicating the presence or absence of past detection in each region of the image.
In FIG. 8B, the detection information history has a 5 × 5 map format, but the screen area dividing method is not limited to this.

9A to 9F, the detection information history of the format shown in FIG. 8B is sequentially updated in the storage unit 103 based on the detection results of the six images F11 to F16 shown in FIG. It is a figure which shows a mode that it goes. The number written in each 5 × 5 area means frequency.
In FIGS. 9A to 9F, since the upper limit value that the frequency can take is 3, the frequency corresponding to the area where the face is detected four times or more remains 3 and is not updated.
In this case, the storage unit 103 stores past detection information in the form of multi-value information indicating the past detection frequency in each region of the image.

  FIG. 10 is a flowchart illustrating an example of an operation flow when the imaging unit 101 is controlled by the control unit 104.

After starting the process, the control unit 104 acquires an exposure evaluation value output from the imaging unit 101 in step ST11.
In step ST <b> 12, the control unit 104 sends control information CTL corresponding to the exposure evaluation value EST from the imaging unit 101 to the imaging unit 101 to control the imaging unit 101.
Next, in step ST13, the control unit 104 reads the detection information history DTCH in the storage unit 103, and determines whether a face has been detected in the past.
If the control unit 104 determines in step ST13 that no face has been detected in the past, the control unit 104 proceeds to the process of step S14 and causes the imaging unit 101 to acquire an exposure evaluation value from the entire screen or the center of the screen. Control and end the process.
On the other hand, when it is determined in step ST13 that the face has been detected in the past, the control unit 104 proceeds to the process of step S15 and acquires the exposure evaluation value only for the region in which the face has been detected in the past. The imaging unit 101 is controlled to end the processing.

According to the operation flow shown in FIG. 10, the exposure evaluation value is acquired only for a region where a face has been detected in the past, that is, a region where the probability that a face exists is high. It is possible to perform exposure control that improves.
However, if a face has not been detected with sufficient frequency in the past, an appropriate exposure evaluation value may not be acquired.
In such a case, it is desirable to perform control so that the exposure evaluation value is acquired from the entire screen or the center of the screen until the ratio of the area in which the face has been detected in the past occupies the predetermined screen or more.

  FIG. 11 is a flowchart illustrating another example of an operation flow when the imaging unit 101 is controlled by the control unit 104.

After starting the process, the control unit 104 initializes the exposure evaluation value of the entire screen to zero in step ST21.
Next, the control part 104 performs the process shown to step ST22-step ST26 shown below with respect to each area | region of a screen.
In step ST22, the control unit 104 acquires partial evaluation values for each area of the screen divided into a plurality of areas from the imaging unit 101, and in subsequent step ST23, the frequency of face detection based on the detection information history DTCH of the storage unit 103. It is determined whether the area is high.
When it is determined in step ST23 that the face detection frequency is high, the control unit 104 updates the partial evaluation value by applying a relatively high weight (weighting) to the partial evaluation value in step ST24. Transition to processing.
On the other hand, if it is determined in step ST23 that the face detection frequency is low, the control unit 104 updates the partial evaluation value by applying a relatively low weight in step ST25, and proceeds to the process of step ST26.
Next, in step ST26, the control unit 104 adds the updated partial evaluation value to the exposure evaluation value for the entire screen.
After completing the processing of step ST22 to step ST26 for all regions, the control unit 104 normalizes the exposure evaluation value according to the weight of each region in step ST27.
And the control part 104 controls the imaging part 101 based on the exposure evaluation value finally obtained in step ST28, and complete | finishes a process.

According to the operation flow shown in FIG. 11, since an exposure evaluation value is calculated by applying a large weight to a region where a face has been detected in the past, that is, a region where the probability that a face exists is high, there is no face. It is possible to perform exposure control that improves the visibility of the area that can be obtained.
However, unlike the operation flow shown in FIG. 10, partial evaluation values are calculated even for areas where no faces existed in the past, so even if the accumulation time of detection results is not sufficient, exceptional processing is performed. There is an advantage that does not need to be done.
In addition, by appropriately setting the weight value for updating the partial evaluation value according to the detection frequency of the face, it is possible to perform exposure control such that the visibility is higher in an area where the probability that a face exists is higher. Become.

  Here, consider the case of shooting for a long time in the day at the same angle as the images shown in FIGS. 2 (a) and 2 (b).

  FIGS. 12A and 12B are diagrams illustrating an example of an image captured at night under the convergence state of the detection information history and exposure control.

Assuming that the format of the detection information history in the storage unit 103 is as shown in FIG. 8B and the upper limit of the frequency is 1 for simplicity, the detection information history converges to the state shown in FIG. Inferred.
That is, a result is obtained that indicates that a face is detected in an area corresponding to the road, and no face is detected in other areas, for example, around street lamps.
Note that the detection rate and detection speed of the detection unit 102 only affect the time taken for the update of the detection information history to converge, and do not cause a large difference in the converged detection information history.
Using this detection information history, according to the operation flow shown in FIG. 11, the weight of the partial evaluation value for the region with the frequency of 1 is high and the weight of the partial evaluation value for the region with the frequency of 0 is low. An evaluation value is calculated. As a result, even when shooting at night from the same angle, exposure control can be performed without being affected by bright areas around the street lamp.
Further, by sacrificing the time taken until the detection information history converges, it is possible to reduce the circuit scale of the detection unit 102 and to reduce power consumption.

FIG. 12B shows an example of an image taken at night under such exposure control.
Compared with the image shown in FIG. 2B, although there is a possibility that the peripheral portion of the streetlight 1 is overexposed, it can be expected that an image maintaining the contrast of an important region where a face can exist can be obtained. In addition, by performing such exposure control, an improvement in the face detection rate itself can be expected.

Further, in the present embodiment, unlike the existing technology, only the weight of the face area detected in a certain image is not increased, but the weight of all areas in which face detection has been performed in the past is increased.
In other words, according to the present embodiment, since the weight for a larger area is increased, it is possible to perform robust control against changes in a small region in the image such as movement of a specific person.
In addition, the total of the areas where face detection has been performed in the past does not change abruptly between successive images, but gradually changes (enlarges) over a long period of time, so a scene in which a person moves In this case, it is possible to avoid a situation where the exposure is unstable.

In the description of the first embodiment, the position where the person exists, and thus the position where the visibility is desired to be improved is specified by the face detection information. However, for example, the position where the person exists is determined by the movement detection information. The exposure control may be performed so as to identify and improve the visibility of the portion.
It is obvious to those skilled in the art that even if the first embodiment is applied and focus control and white balance control are performed instead of exposure control, the visibility of a region where a face can exist can be improved. .

<Second Embodiment>
In the second embodiment and the third embodiment to be described later, a case where the present invention is applied to quasi-fixed point shooting, that is, a case of shooting at a different angle depending on a time zone will be described as an example. .

FIG. 13 is a block diagram illustrating an example of a configuration of an imaging apparatus (camera apparatus) to which the image processing apparatus according to the second embodiment of the present invention is applied.
In the second embodiment, blocks having the same configuration as those of the imaging apparatus (image processing apparatus) of the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  As illustrated in FIG. 13, the imaging apparatus 100A according to the second embodiment includes an imaging unit 101A, a detection unit 102, a storage unit 103A, a control unit 104, and an erasing unit 105.

The imaging unit 101A receives a light wave composed of reflected light from a subject, converts the light wave into an image, and outputs the image to a processing system such as a detection unit 102 at a later stage.
This configuration is almost the same as that of the imaging unit 101 in the imaging apparatus 100 according to the first embodiment of the present invention. The imaging unit 101A of the second embodiment includes a motor (not shown) for changing the shooting angle, and outputs the angle information ANGL that is information of the current shooting angle. The imaging unit of the first embodiment. 101.
The imaging unit 101A outputs the angle information ANGL to the erasing unit 105.

The storage unit 103A receives the detection information DTC from the detection unit 102, converts it into a predetermined format, and stores it.
This configuration is almost the same as that of the storage unit 103 in the imaging apparatus 100 according to the first embodiment of the present invention. The storage unit 103A according to the second embodiment is different from the storage unit 103 according to the first embodiment in that when the erasure signal ERS is received from the outside, all or part of the detection information history DTCH is discarded.

  When the erasing unit 105 receives the angle information ANGL from the imaging unit 101A and detects a change in the shooting angle, the erasing unit 105 outputs an erasing signal ERS to the storage unit 103A.

  FIGS. 14A to 14E are diagrams for explaining an example of application of the second embodiment.

FIG. 14A is an image diagram of a dome-type surveillance camera 200 that can change the shooting angle by a built-in motor.
The dome-type monitoring camera 200 is an example of an application of the imaging device in the second embodiment.
FIG. 14B is an image diagram when the L-shaped indoor corridor 210 is viewed from directly above.
It is assumed that a dome-type surveillance camera 200 shown in FIG. 14A is installed on the ceiling at a position corresponding to the black circle portion in the figure.

FIG. 14C-1 is an image diagram illustrating an example of a photographed image when the photographing angle is set in the direction of the right arrow RA (angle A) in FIG.
FIG. 14C-2 is an image diagram illustrating an example of a photographed image when the photographing angle is set in the direction of the upward arrow UA (angle B) in FIG.
As shown in FIGS. 14 (c-1) and (c-2), when two (or more) shooting angles are switched in a time zone or the like, it is inconvenient to use the imaging device of the first embodiment. May occur.

FIGS. 14 (d-1) and 14 (d-2) are images taken for a long time using the imaging apparatus 100 according to the first embodiment of the present invention in a situation where the shooting angle is fixed to the angle A and the angle B, respectively. It is a figure which shows an example of the detection information log | history at the time of performing.
The area where the face can exist on the screen varies depending on the shooting angle.

FIG. 14E is a diagram illustrating an example of the detection information history when the two shooting angles are appropriately switched and long-time shooting is performed using the imaging apparatus 100 according to the first embodiment of the present invention.
The detection information history shown in FIG. 14 (e) is equivalent to a combination of the detection information histories shown in FIGS. 14 (d-1) and (d-2).
For this reason, exposure control is performed on the assumption that a face is detected even in a region where a face cannot exist at angle A, or a region where a face cannot exist at angle B. This is not the expected control.

In order to solve such a problem, the second embodiment further includes an erasing unit 105 in addition to the configuration of the first embodiment.
Hereinafter, the operation of the erasing unit 105 will be mainly described.

  FIG. 15 is a flowchart showing an operation flow in which the erasure unit 105 discards the detection information history in the storage unit 103A.

After starting the processing, in step ST31, the erasure unit 105 refers to the angle information ANGL output from the imaging unit 101A, and determines whether or not there has been a change from the shooting angle at the previous reference.
If the erasing unit 105 determines in step ST31 that the shooting angle has not changed, the erasing unit 105 ends the process without doing anything.
On the other hand, when determining that the shooting angle has changed in step ST31, the erasure unit 105 discards all of the detection information history DTCH in the storage unit 103A in step ST32 and ends the process.

For example, of the 24 hours a day, 12 hours in the morning are fixed at angle A, and 12 hours in the afternoon are fixed at angle B.
Further, when the switching of the shooting angle does not occur, the detection information history in the accumulation unit 103A is sequentially updated in the operation flow shown in FIG.
Then, immediately after the angle is switched, the erasure unit 105 once discards the detection information history in the storage unit 103A.
When a sufficient time has elapsed after the angle is switched, the detection information history in the storage unit 103A is asymptotic to that shown in FIG. 14 (d-1) in the morning and that shown in FIG. 14 (d-2) in the afternoon. However, it does not become the one shown in FIG.
However, also in the second embodiment, the format shown in FIG. 8B is adopted as the format of the detection information history as in the description of the first embodiment.
That is, the problem in applying the imaging apparatus according to the first embodiment of the present invention is solved.

When the shooting angle of the imaging unit 101A is determined by the time zone, the erasing unit 105 operates not to change the angle information but to discard the detection information history triggered by a time change such as AM to PM and PM to AM. However, the same effect can be obtained.
In addition, even when the shooting angle is completely fixed, if the area where a person can exist differs depending on the time zone, an erasure unit is added to the imaging apparatus according to the first embodiment, and the predetermined time is set. Then, the detection information history in the storage unit 103 may be discarded.

Note that when applied to shooting at a fixed angle, the erasure unit 105 does not necessarily need to discard the entire detection information history.
For example, when the format shown in FIG. 7B is used as the format of the detection information history, only the oldest detection information in the detection information history is discarded when erasing.
Thereby, it is possible to avoid an abrupt change in the content of the detection information history due to the discarding, and to prevent the brightness of the image from changing greatly before and after the discarding.
Alternatively, the new detection information history from the present to the predetermined time T before is stored in the format shown in FIG. 7B, and the old detection information history before the predetermined time T is stored in the format shown in FIG. 8B. It doesn't matter.
By holding the detection information history in such a hybrid format, even when the capacity of the storage unit 103A is limited, it is possible to hold the detection information history over a long period of time. The detection information older than that can be discarded.

<Third Embodiment>
The imaging apparatus 100B according to the third embodiment assumes a case where a plurality of imaging angles are switched by time or the like as in the imaging apparatus 100A according to the second embodiment described above.
For the sake of simplicity of explanation, here, it is assumed that detection information is accumulated by switching between two types of shooting angles, angle A and angle B, in the second embodiment.

FIG. 16 is a block diagram illustrating an example of a configuration of an imaging apparatus (camera apparatus) to which the image processing apparatus according to the third embodiment of the present invention is applied.
Note that in the third embodiment, blocks having the same configuration as those of the imaging devices (image processing devices) described in the first and second embodiments are denoted by the same reference numerals, and detailed description thereof is omitted. To do.

  As illustrated in FIG. 16, the imaging apparatus 100B according to the third embodiment includes an imaging unit 101A, a detection unit 102, storage units 103B-1 and 103B-2, a control unit 104, switches 106-1 and 106-2, And a switching unit 107.

Assume that storage unit 103B-1 stores detection information DTC-A corresponding to angle A, and storage unit 103B-2 stores detection information DTC-B corresponding to angle B.
The accumulating units 103B-1 and 103B-2 are denoted by the same reference numerals as the accumulating unit 103, although different reference numerals are attached for convenience of explanation.

The switch 106-1 selects whether the detection information DTC output from the detection unit 102 is stored in the storage unit 103B-1 or the storage unit 103B-2 based on the switching signal SSW output from the switching unit 107. .
In the switch 106-1, the terminal a is connected to the output line of the detection information DTC of the detection unit 102, the terminal b is connected to the input unit of the storage unit 103B-1, and the terminal c is connected to the input unit of the storage unit 103B-2. It is connected.
The switch 106-1 selectively connects the terminal a with the terminal b or the terminal c according to the switching signal SSW.

The switch 106-2 selects, based on the switching signal SSW output from the switching unit 107, whether the detection information history DTCH input to the control unit 104 is acquired from the storage unit 103B-1 or the storage unit 103B-2. To do.
In the switch 106-2, the terminal a is connected to the detection information history input unit of the control unit 104, the terminal b is connected to the output unit of the storage unit 103B-1, and the terminal c is connected to the output unit of the storage unit 103B-2. It is connected.
The switch 106-2 selectively connects the terminal a with the terminal b or the terminal c according to the switching signal SSW.

  When the switching unit 107 receives the angle information ANGL from the imaging unit 101A and detects a change in the shooting angle, the switching unit 107 outputs a switching signal SSW corresponding to the current shooting angle to the switches 106-1 and 106-2.

  FIG. 17 is a flowchart illustrating an operation flow in which the switching unit 107 outputs a switching signal for switching between the switch 106-1 and the switch 106-2.

After starting the processing, the switching unit 107 refers to the angle information ANGL output from the imaging unit 101A in step ST41 and determines whether or not the current shooting angle is the angle A.
If the switching unit 107 determines that the angle is A in step ST41, the switching unit 107 outputs a switching signal SSW to the switches 106-1 and 106-2 to connect to the storage unit 103B-1 corresponding to the angle A in step ST42. Then, the process ends.
On the other hand, if the switching unit 107 determines that it is not the angle A in step ST41, that is, if it is determined that the angle is B, the switch 106- is connected to the storage unit 103B-2 corresponding to the angle B in step ST43. The switching signal SSW is output to 1, 106-2, and the process is terminated.

In the second embodiment, when the angle is switched, the only detection information history in the storage unit 103A is once discarded.
Therefore, in the case where the angle switching time interval is short, the detection information is not sufficiently accumulated, and there is a possibility that the effect cannot be exhibited sufficiently.
However, in the third embodiment, by providing the storage units as many as the number of angles, the detection information stored at the time of switching the angles is not discarded, so that the above problem can be solved.

  In the description of the third embodiment, for the sake of simplicity, the shooting angles are limited to two types, angle A and angle B. However, the storage unit has the same number of angles, and the switching unit corresponds to the angle. It goes without saying that if the switch is controlled so as to switch to the storage unit, it is possible to cope with three or more angles.

<Fourth Embodiment>
In the fourth embodiment, an improved version of the imaging apparatus according to the first embodiment is presented.

FIG. 18 is a block diagram illustrating an example of a configuration of an imaging apparatus (camera apparatus) to which the image processing apparatus according to the fourth embodiment of the present invention is applied.
Note that in the fourth embodiment, blocks having the same configuration as those of the imaging device (image processing device) described in the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.

  As illustrated in FIG. 18, the imaging apparatus 100C of the third embodiment includes an imaging unit 101, a detection unit 102C, a storage unit 103, a control unit 104, a convergence degree acquisition unit 108, and a range setting unit 109.

The detection unit 102C receives an image, detects a human face from the image, and outputs detection information DTCV to the storage unit 103 and the convergence degree acquisition unit 108.
This configuration is substantially the same as that of the detection unit 102 in the imaging apparatus 100 according to the first embodiment of the present invention. The first embodiment is that the detection unit 102C according to the fourth embodiment can input range information RNG in addition to an image, and can specify from the outside which range in the image is a target for face detection. Different from the detection unit 102 of the embodiment.

The convergence degree acquisition unit 108 inputs the detection information DTC from the detection unit 102C and the detection information history DTCH from the storage unit 103, and updates the history convergence degree.
The convergence degree acquisition unit 108 outputs the history convergence degree to the range setting unit 109.
When shooting for a long time with a fixed shooting angle, the area where the face is detected gradually expands immediately after the start of shooting, but no new area is detected over time. It converges to the situation.
Such a situation is called that the detection information history converges.
The history convergence degree is a numerical value indicating how close the detection information history DTCH is to the convergence state, and the larger the value, the closer to the convergence state.
Note that the storage area for holding the history convergence is omitted in the block diagram shown in FIG. The history convergence is assumed to be initialized to zero before the imaging apparatus 100C according to the fourth embodiment starts operation.

The range setting unit 109 inputs the detection information history DTCH by the storage unit 103 and the history convergence level HCNV by the convergence level acquisition unit 108, and sets and sets a detection range excluding the area where no face is detected The range information RNG is output to the detection unit 102C.
The range information is information for indicating to which range of the input image the detection unit 102C performs face detection.

  FIG. 19 is a flowchart showing an operation flow in which the convergence degree acquisition unit 108 updates the history convergence degree.

After the start of processing, the convergence degree acquisition unit 108 determines whether or not one or more faces have been detected in step ST51 with reference to the current detection information.
If the convergence degree obtaining unit 108 determines in step ST51 that no face has been detected, the convergence degree obtaining unit 108 ends the process.
On the other hand, when it is determined in step ST51 that at least one face has been detected, the convergence degree acquiring unit 108 includes the face area detected this time in the area in which the face has been detected in the past in next step ST52. It is determined whether or not.
It should be noted that the convergence level acquisition unit 108 can know in which area the face has been detected in the past by referring to the detection information history DTCH.
If it is determined that the convergence is not included in step ST52, the convergence acquisition unit 108 decreases the convergence by a predetermined value in step ST53, and ends the process.
On the other hand, when it is determined that the convergence degree acquisition unit 108 is included in step ST52, the convergence degree is increased by a predetermined value in step 54, and the process ends.

  FIG. 20 is a flowchart illustrating an operation flow in which the range setting unit 109 generates range information.

After the start of processing, range setting section 109 determines in step ST61 whether or not the input history convergence degree HCNV value is equal to or greater than a predetermined threshold value.
If the range setting unit 109 determines in step ST61 that the degree of convergence is less than the predetermined threshold value, the range setting unit 109 regards that the detection information history has not converged, and outputs the entire input image as range information RNG in step ST62. finish.
On the other hand, if the range setting unit 109 determines in step ST61 that the degree of convergence is greater than or equal to a predetermined threshold, the range setting unit 109 regards the detection information history as converged and refers to the detection information history.
Then, the range setting unit 109 generates and outputs range information including a region where the face has been detected at least N times in the past (N is a positive integer) in step ST63, and ends the processing. .
Here, when the format of the detection information history is as shown in FIG. 8B, the maximum value that N can take is equivalent to the upper limit value of the frequency.
The predetermined method is specifically a method of generating range information in the range of the area in which the face has been detected in the past, or range information in the range of a rectangle including the area in which the face has been detected in the past. The method of producing | generating etc. is mentioned. However, any other method using the detection information history may be used.
If face detection is not used for purposes other than those described in the embodiment of the present invention, the range may be fixedly set to zero after the detection information history has converged, and face detection may not be performed in practice.

Here, if the convergence is set to zero in step ST53 of the flowchart shown in FIG. 19, the following processing is performed.
If the newly detected face area is within the previously detected face area for a predetermined number of times, the detection information history is considered to have converged in step S61 of the flowchart shown in FIG. The detection target range is limited to the area where the face has been detected in the past.

According to the imaging apparatus according to the first embodiment, the entire input image is always the target for face detection.
On the other hand, in the imaging device according to the fourth embodiment, after the detection information history converges, face detection is performed by excluding an area where no face has been detected so far. Although it depends on the face detection method, it is obvious that the cost for face detection is generally reduced when the area to be detected is narrow. That is, power consumption for face detection can be reduced.
However, when the convergence of the detection information history is erroneously determined, there is a possibility that the area where the face actually exists is excluded from detection.

  FIG. 21 is a diagram illustrating an example of a graph in which the horizontal axis indicates the number of face detections and the vertical axis indicates the number of face detection areas in the past.

Note that the graph shown in FIG. 21 always assumes that the entire input image is the face detection target. In the period D01, since the area is newly increased every time face detection is performed, the detection information history is not regarded as convergence.
However, in the period D02, even if the face is detected, all are within the area where the face has been detected in the past. Therefore, if the predetermined number of times is, for example, five, the detection information history has converged during the period D02. Will be considered.
However, in practice, a face is detected again in a new area in the period D03.
Therefore, it is assumed that the range setting unit 109 is operated in the period D03 based on the flowchart shown in FIG. Then, the range to be a face detection target is limited to a region where a face has been detected in the past at the time period D02, and a region that should be newly detected in the time period D03 is excluded from the face detection target. The face is detected. As a result, face detection leaks occur after the period D03.

  In order to avoid such a situation, the range setting unit 109 outputs the range information so that the entire input image is detected once every time face detection is performed a predetermined number of times even after the convergence of the detection information history. It is desirable to correct the behavior of

As described above, according to the present embodiment, the following effects can be obtained.
In moving image shooting at a fixed point or a quasi-fixed point, it is possible to improve the visibility of an area where a human face is highly likely to exist.
By using face detection information detected under lighting conditions favorable to the face detection rate, even when shooting under lighting conditions unfavorable to the face detection rate, the visibility of areas where there is a high probability that a face exists is improved. In addition, the face detection rate can be improved.
Since faces that could not be detected in the past can be detected, the applicable scenes can be expanded even when face detection is used for another purpose.
After operating for a long time, power consumption for face detection can be achieved by excluding the area where the face does not exist from the subsequent face detection target range, or by setting the detection target range to zero and stopping the face detection operation itself. Can be reduced.
Based on the detection information history accumulated for a long time, exposure control (or focus control, white balance control) is performed. This can be realized, and flicker in moving images can be reduced.

Note that the method described above in detail can be formed as a program according to the above-described procedure and executed by a computer such as a CPU.
Such a program can be configured to be accessed by a recording medium such as a semiconductor memory, a magnetic disk, an optical disk, a floppy (registered trademark) disk, or the like, and to execute the program by a computer in which the recording medium is set.

It is a figure which shows an example of the input image in a certain time and the time when it passed. It is a figure which shows an example of the image image | photographed in the daytime and the night with the same angle. It is a figure which shows the frame group which comprises an example of the scene where a person appears. 1 is a block diagram illustrating an example of a configuration of an imaging apparatus (camera apparatus) to which an image processing apparatus according to a first embodiment of the present invention is applied. It is a block diagram which shows an example of a detailed structure of the imaging part which concerns on this embodiment. It is a flowchart which shows an example of the operation | movement flow until it inputs a light wave and builds a detection information log | history in a storage part. It is a figure which shows an example of the state of the face detection in this embodiment, and the format of a detection information log | history. It is a figure which shows another example of the condition of the face detection in this embodiment, and the format of a detection information log | history. It is a figure which shows a mode that the detection information log | history of the format shown in FIG.8 (b) is sequentially updated in the accumulation | storage part by the detection result of six images shown to Fig.8 (a). It is a flowchart which shows an example of the operation | movement flow at the time of controlling an imaging part by a control part. It is a flowchart which shows another example of the operation | movement flow at the time of controlling an imaging part by a control part. It is a figure which shows an example of the image image | photographed at night under the convergence state of detection information history, and exposure control. It is a block diagram which shows an example of a structure of the imaging device (camera apparatus) to which the image processing apparatus which concerns on the 2nd Embodiment of this invention is applied. It is a figure for demonstrating an example of the application of the 2nd embodiment. It is a flowchart which shows the operation | movement flow which an erasure | elimination part discards the detection information log | history in a storage part. It is a block diagram which shows an example of a structure of the imaging device (camera device) to which the image processing apparatus which concerns on the 3rd Embodiment of this invention is applied. It is a flowchart which shows the operation | movement flow which a switching part outputs the switching signal which switches two switches. It is a block diagram which shows an example of a structure of the imaging device (camera device) to which the image processing apparatus which concerns on the 4th Embodiment of this invention is applied. It is a flowchart which shows the operation | movement flow in which a convergence degree acquisition part updates a history convergence degree. It is a flowchart which shows the operation | movement flow in which a range setting part produces | generates range information. It is a figure which shows an example of the graph which plotted the frequency | count of face detection on the horizontal axis, and plotted the number of the area | regions where the vertical axis | shaft was detected in the past.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100,100A-100C ... Imaging device (camera apparatus), 101, 101A ... Imaging part, 1011 ... Lens, 1012 ... Aperture, 1013 ... Imaging element, 1014 ... Analog-digital (AD) converter, 1015 ... exposure evaluation value calculation unit, 102 ... detection unit, 103, 103A, 103B-1, 103B-2 ... storage unit, 104 ... control unit, 105 ... Erasing unit, 106-1, 106-2 ... switch, 107 ... switching unit, 108 ... convergence degree obtaining unit, 109 ... range setting unit.

Claims (26)

A detection unit that detects a moving region in the screen from an image captured in a state according to control information, and outputs detection information;
An accumulation unit that accumulates past detection information output by the detection unit;
An image processing apparatus comprising: a control unit that generates the control information based on past detection information stored in the storage unit and controls an imaging state according to the control information. The controller is
It is determined whether or not a moving area has been detected in the past based on the detection information history by the storage unit. If it is determined that no moving area has been detected in the past, an evaluation value is acquired from the entire screen or the center of the screen. When it is determined that a moving area has been detected in the past, the control information is generated so as to obtain the evaluation value for the area in which the moving area has been detected in the past,
The image processing apparatus according to claim 1, wherein the imaging condition is changed by control information corresponding to the evaluation value. The controller is
The image processing apparatus according to claim 2, wherein control is performed so that an evaluation value is acquired from the entire screen or the center of the screen until the ratio of the area in which the moving area has been detected in the past occupies a predetermined value or more. The controller is
Get partial evaluation values for each area of the screen divided into multiple areas,
It is determined whether each region from which the partial evaluation value is acquired is a region where the detection frequency of the storage unit is high based on the detection information history of the storage unit. If the partial evaluation value is updated with relatively high weighting and the detection frequency is determined to be low, the partial evaluation value is updated with relatively low weighting, and the updated partial evaluation value is taken into account. The image processing apparatus according to claim 1, wherein an evaluation value for the entire screen is obtained and imaging conditions are changed by control information corresponding to the evaluation value for the entire screen. The controller is
The image processing apparatus according to claim 4, wherein a weighting value for updating the partial evaluation value is set according to a detection frequency of a moving area. The image processing apparatus according to claim 1, further comprising an erasing unit that erases part or all of past detection information accumulated in the accumulation unit under a predetermined condition. The predetermined condition is
The image processing apparatus according to claim 6, wherein a predetermined time has elapsed since the previous erasure. In the image,
Includes images taken with different shooting angles,
The eraser is
When a change in shooting angle is detected from the angle information, an erasure signal is output to the storage unit,
The storage unit
The image processing apparatus according to claim 6, wherein in response to the erasure signal, at least a part of the accumulated past detection information is discarded. It has a switching unit that performs switching processing according to angle information,
In the image,
Includes images taken with different shooting angles,
The storage unit
Including N independent storage units corresponding to the N kinds of angles (N is a natural number),
The switching unit is
The image processing apparatus according to any one of claims 1 to 5, wherein the storage unit corresponding to the angle information is switched to store corresponding detection information and input the detection information history to the control unit. Based on the detection information by the detection unit and the detection information history of the storage unit, a convergence level acquisition unit that acquires the convergence level of past detection information history stored in the storage unit;
A range setting unit that generates range information indicating a range of an input image that is a target of detection of a moving region from the convergence degree and the detection information history,
The detector is
The image processing apparatus according to claim 1, wherein the moving region is detected within a range set by a range setting unit. The convergence degree acquisition unit
If it is recognized that at least one moving area has been detected by referring to the detection information by the detection unit, whether the moving area detected from the detection information history of the storage unit is included in the area where the moving area has been detected in the past The image according to claim 10, wherein when it is determined that it is not included, the convergence is decreased, and when it is determined that it is included, the convergence is increased and output to the range setting unit. Processing equipment. The range setting unit includes:
The image processing apparatus according to claim 10 or 11, wherein a range is set so that the entire screen is a detection target every time a moving area is detected a predetermined number of times after convergence of the detection information history. At least an imaging unit that captures an image based on external control;
A detection unit that detects a moving area in a screen from an image captured by the imaging unit and outputs detection information;
An accumulation unit that accumulates past detection information output by the detection unit;
An imaging apparatus comprising: a control unit that performs predetermined control on the imaging unit based on past detection information history accumulated in the accumulation unit. The imaging unit
Obtaining an evaluation value related to imaging of an image and outputting it to the control unit, and having a function of changing imaging conditions according to control information of the control unit,
The controller is
It is determined whether or not a moving area has been detected in the past based on the detection information history by the storage unit. If it is determined that no moving area has been detected in the past, the evaluation value is acquired from the entire screen or the center of the screen. If it is determined that a moving area has been detected in the past, the imaging unit is controlled to acquire the evaluation value for the area in which the moving area has been detected in the past,
The imaging apparatus according to claim 13, wherein control information corresponding to an evaluation value by the imaging unit is output to the imaging unit to change imaging conditions. The controller is
The imaging apparatus according to claim 14, wherein control is performed so that an evaluation value is acquired from the entire screen or the center of the screen until the ratio of the area in which the moving area has been detected in the past occupies a predetermined ratio or more. The imaging unit
Obtaining an evaluation value related to imaging of an image and outputting it to the control unit, and having a function of changing imaging conditions according to control information of the control unit,
The controller is
Control to acquire a partial evaluation value for each area of the screen divided into a plurality of areas in the imaging unit,
It is determined whether each region from which the partial evaluation value is acquired is a region where the detection frequency of the storage unit is high based on the detection information history of the storage unit. If the partial evaluation value is updated with relatively high weighting and the detection frequency is determined to be low, the partial evaluation value is updated with relatively low weighting, and the updated partial evaluation value is taken into account. The imaging apparatus according to claim 13, wherein an evaluation value for the entire screen is obtained, and control information corresponding to the evaluation value for the entire screen is output to the imaging unit to change imaging conditions. The controller is
The imaging apparatus according to claim 16, wherein a weighting value for updating the partial evaluation value is set according to a detection frequency of a moving area. The imaging device according to any one of claims 13 to 17, further comprising an erasing unit that erases part or all of past detection information accumulated in the accumulation unit under a predetermined condition. The predetermined condition is
The imaging apparatus according to claim 18, wherein a predetermined time has elapsed since the previous erasure. The imaging unit
It is possible to take an image by changing the shooting angle, and output the current angle information to the erasure unit,
The eraser is
When a change in shooting angle is detected from the angle information, an erasure signal is output to the storage unit,
The storage unit
The imaging apparatus according to claim 18, wherein in response to the erasure signal, at least a part of the accumulated past detection information is discarded. It has a switching unit that performs switching processing according to angle information,
The imaging unit
It is possible to take an image by changing the shooting angle, and output the current angle information to the switching unit,
The storage unit
Including N independent storage units corresponding to the N kinds of angles (N is a natural number),
The switching unit is
The imaging device according to any one of claims 13 to 17, wherein the storage unit corresponding to the angle information is switched to store corresponding detection information and input the detection information history to the control unit. Based on the detection information by the detection unit and the detection information history of the storage unit, a convergence level acquisition unit that acquires the convergence level of past detection information history stored in the storage unit;
A range setting unit that generates range information indicating a range of an input image that is a target of detection of a moving region from the convergence degree and the detection information history,
The detector is
The imaging device according to any one of claims 13 to 21, wherein a moving region is detected within a range set by a range setting unit. The convergence degree acquisition unit
If it is recognized that at least one moving area has been detected by referring to the detection information by the detection unit, whether the moving area detected from the detection information history of the storage unit is included in the area where the moving area has been detected in the past The degree of convergence is decreased when it is determined that it is not included, and the degree of convergence is increased when it is determined that it is included, and is output to the range setting unit. apparatus. The range setting unit includes:
The imaging device according to claim 22 or 23, wherein a range is set so that the entire screen is a detection target every time a moving area is detected a predetermined number of times after convergence of the detection information history. A detection step of detecting a moving region in the screen from an image captured in a state corresponding to the control information and obtaining detection information;
An accumulation step of accumulating past detection information obtained in the detection step;
A generation step of generating the control information based on the past detection information accumulated in the accumulation step;
An image processing method comprising: a control step of controlling an imaging state in accordance with the control information. A detection process for detecting a moving area in the screen from an image captured in a state corresponding to the control information and obtaining detection information;
An accumulation process for accumulating past detection information obtained by the detection process;
A generation process for generating the control information based on past detection information accumulated in the accumulation process;
A program that causes a computer to execute image processing including control processing that controls an imaging state according to the control information.

Images