JP2012226583A - Image processor, image processing system and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce recognition processing.SOLUTION: The image processor includes: acquisition means for acquiring an image picked up by first image pickup means; detection means for detecting an object area from the image acquired by the acquisition means; determination means for, when there exists the object area detected by the detection means in a shared area between the first image pickup means and second image pickup means in the image, determining whether to execute the object recognition processing of the shared area by an image processor or a second image processor; setting means for, when it is determined that the object recognition processing should be executed by the image processor by the determination means, setting an object recognition range to the whole range of the image, and for, when it is determined that the object recognition processing should be executed by the second image processor by the determination means, setting the object recognition range to an area obtained by excluding the shared area from the whole range of the image; and recognition means for executing object recognition to determine whether or not an object is projected in the object recognition range set by the setting means.

Description

  The present invention relates to an image processing apparatus, an image processing system, and an image processing method.

2. Description of the Related Art Conventionally, cameras have been installed for surveillance purposes in stores and building ceilings, streets, factories, and the like for the purpose of improving security such as crime prevention. In addition, systems have been proposed that use surveillance cameras to count people passing through convenience stores, shopping malls, supermarkets, etc., and public places, analyze human flow, and use surveillance cameras for marketing purposes. .
A technique for detecting a subject such as a face in such an image has been proposed in Non-Patent Document 1 shown below. In Non-Patent Document 1, a face pattern in an image is detected by a neural network. The operation of the apparatus using the technique of Non-Patent Document 1 will be briefly described below. First, the apparatus reads an image targeted for face detection into a memory, and cuts out a predetermined area to be compared with the face from the read image. Then, the apparatus receives a distribution of pixel values of the cut out region as an input, and obtains one output by calculation using a neural network. At this time, the weights and thresholds of the neural network are learned in advance using face image patterns and non-face image patterns, and, for example, the device is a face if the output of the neural network is 0 or more, and is non-face otherwise It is determined that Then, the apparatus detects the face from the image by scanning the cutout position of the image pattern to be collated with the face which is an input of the neural network in the vertical and horizontal directions from the entire image. Further, in order to cope with detection of faces of various sizes, the apparatus sequentially reduces the image at a predetermined ratio and performs the above-described face detection scanning.
Note that Non-Patent Document 1 describes a technique for detecting a face from an image, but the same can be realized when detecting other objects from an image.

Patent Document 1 is disclosed as a technique for recognizing an object from an image captured by a camera. The technique of Patent Document 1 is a technique in which a background difference is first performed to narrow down a region that has changed from a background image, a moving region is detected, and then that region is evaluated in detail to identify individual objects.
On the other hand, with the enlargement of the imaging range, a system for installing a plurality of cameras and simultaneously monitoring and photographing them has been put into practical use. However, a considerable amount of processing performance is required to simultaneously perform face detection, object recognition processing, and the like as described above on a plurality of camera images. Therefore, there is a problem of how to reduce the processing load. As a technique for solving this problem, for example, the technique of Patent Document 2 has been proposed. The technique of Patent Document 2 is a technique in which a specific person to be detected is marked and the subsequent detection process is terminated at the timing when the person is detected from an image. Further, Japanese Patent Laid-Open No. 2004-228867 proposes a technique in which a specific object in all images is detected and counted, only an optimal image is selected based on the count, transmitted to a server, and then processed.

JP 2003-157440 A JP 2007-243270 A Japanese Patent Laid-Open No. 2005-347942

Rowley et al, “Neural network-based face detection”, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 20, NO. 1, JANUARY 1998

  However, the above-described technique is limited to the monitoring application of a specific person, or the image to be recognized is limited. For example, when the above-described technique is applied when it is necessary to perform recognition processing for all moving objects in all images, there is a problem that a load is applied to the recognition processing and frames are dropped in the recognition processing.

  The present invention has been made in view of such problems, and an object thereof is to reduce recognition processing.

  Therefore, the present invention performs image processing of an image picked up by the first image pickup means, and performs image processing of an image picked up by the second image pickup means that partially shares the image pickup range with the first image pickup means. An image processing apparatus that communicates with a second image processing apparatus to perform, an acquisition unit that acquires an image captured by the first imaging unit, and an object region that is detected from the image acquired by the acquisition unit When the object area detected by the detection means is in a shared area between the detection means and the first imaging means and the second imaging means in the acquired image, the object recognition processing of the shared area A determination means for determining whether the image processing apparatus or the second image processing apparatus performs the image recognition, and the determination means determines that the image recognition apparatus performs the object recognition range for the entire image. Set to the range, If the second image processing apparatus determines that the object recognition range is to be set, the object recognition range is set to an area excluding the shared area from the entire range of the image, and the object recognition range set by the setting means Recognizing means for recognizing whether an object is captured or not.

  According to the present invention, recognition processing can be reduced.

It is a figure which shows an example of the hardware constitutions of an image processing apparatus. 1 is a diagram illustrating an example of a system configuration and the like of an image processing system according to a first embodiment. 3 is a flowchart illustrating an overview of the first embodiment. It is a flowchart which shows an example of the detailed process of step S100. 6 is a diagram illustrating an example of an image acquired by an image acquisition unit 101 when a person is imaged by the imaging device 110. FIG. It is a figure which shows an example of the background image used as the reference | standard model memorize | stored in the memory | storage device. It is a figure which shows an example of a reference | standard model. It is a figure which shows an example of the image which shows the moving body area | region which the object area | region detection part 102 extracted. FIG. 6 is a diagram illustrating an example of an image acquired by an image acquisition unit 201 when the image capturing apparatus 210 captures an image at the same time as in FIG. 5. It is a figure which shows an example of the result obtained when step S104 is performed in FIG. It is a flowchart which shows an example of the detailed process of step S200. FIG. 4 is a diagram illustrating an example of a system configuration and the like of an image processing system according to a second embodiment. 10 is a flowchart illustrating an overview of the second embodiment. It is a flowchart which shows an example of the detailed process of step S400. 5 is a flowchart illustrating an example of processing of the image processing apparatus 300. 5 is a flowchart illustrating an example of processing of the image processing apparatus 400.

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

<Embodiment 1>
FIG. 1 is a diagram illustrating an example of a hardware configuration of the image processing apparatus.
As shown in FIG. 1, the image processing apparatus 100 includes a control device 10, a storage device 11, and a network interface device 12 as a hardware configuration. The control device 10 is a CPU or the like, and executes processing based on a program stored in the storage device 11, thereby realizing a software configuration of the image processing device 100 described later and processing of a flowchart. The storage device 11 is a ROM, RAM, HD, or the like, and stores a program or an image. The network interface device 12 connects the image processing device 100 to a network.
The image processing apparatus 200 has the same hardware configuration as the image processing apparatus 100.
In the present embodiment, the function of the image processing apparatus in FIG. 2 described later is described as being implemented as software. However, each function of the image processing apparatus in FIG. You may mount in an apparatus. In the present embodiment, as will be described later, the imaging apparatus and the image processing apparatus are described as separate units, but they may be in the same apparatus.

FIG. 2 is a diagram illustrating an example of a system configuration and the like of the image processing system according to the first embodiment.
As shown in FIG. 2, the image processing system includes imaging devices 110 and 210 that partially share a visual field or imaging range, and image processing devices 100 and 200 that detect a person from images captured by the imaging device. ,including. Here, the imaging device 110 is an example of a first imaging unit. The imaging device 210 is an example of a second imaging unit. The image processing apparatus 100 is an example of a first image processing apparatus. The image processing apparatus 200 is an example of a second image processing apparatus.
The imaging device 110 is a device that captures an image, such as a video camera, and includes an imaging lens and an imaging element that performs photoelectric conversion to generate a pixel signal. The captured image is input to the image acquisition unit 101.
The image processing apparatus 100 is an image processing apparatus proposed in this embodiment.
The image acquisition unit 101 acquires an image captured by the imaging device 110.
The object area detection unit 102 extracts a specific image area from the image acquired by the image acquisition unit 101.
In the present embodiment, the object area detection unit 102 extracts a moving object area from the image acquired by the image acquisition unit 101.
The load estimation unit 103 estimates a processing load when the object recognition unit 105 performs a recognition process from the image region extracted by the object region detection unit 102.
The recognition area limiting unit 104 compares the processing load amount estimated by the load estimation unit 103 with the processing load amount received from the communication unit 106, and limits the region in which the object recognition unit 105 performs the recognition process.
The object recognition unit 105 recognizes and detects whether or not a desired object is shown in the specific image region extracted by the object region detection unit 102.
In the present embodiment, the object recognition unit 105 extracts a human body region from the specific image region extracted by the object region detection unit 102.
The communication unit 106 transmits and receives a processing load amount with the communication unit 206 of the image processing apparatus 200.

Similarly, the imaging device 210 is a device that captures an image, such as a video camera, and includes an imaging lens and an imaging device that performs photoelectric conversion to generate a pixel signal. The captured image is input to the image acquisition unit 201.
The image processing apparatus 200 is an image processing apparatus proposed in this embodiment.
The image acquisition unit 201 acquires an image captured by the imaging device 210.
The object area detection unit 202 extracts a specific image area from the image acquired by the image acquisition unit 201.
In the present embodiment, the object area detection unit 202 extracts a moving object area from the image acquired by the image acquisition unit 201.
The load estimation unit 203 estimates the processing load when the object recognition unit 205 performs the recognition process from the image region extracted by the object region detection unit 202.
The recognition area limiting unit 204 compares the processing load amount estimated by the load estimation unit 203 with the processing load amount received from the communication unit 206, and limits the region where the object recognition unit 205 performs the recognition process.
The object recognition unit 205 recognizes and detects whether or not a desired object is shown in the specific image region extracted by the object region detection unit 202.
In the present embodiment, the object recognition unit 205 extracts a human body region from the specific image region extracted by the object region detection unit 202.
The communication unit 206 transmits and receives processing load amounts to and from the communication unit 106 of the image processing apparatus 100.

FIG. 3 is a flowchart showing an outline of the first embodiment. The following will be described in order. The operation of the following flowchart will be described for the image processing apparatus 100, but the image processing apparatus 200 also performs the same operation.
Step S100:
First, the object area detection unit 102 extracts a moving object area from the image acquired by the image acquisition unit 101. Here, the moving object area is an area that can be extracted by the object area detecting unit 102 taking a difference from an image of only the background having the same angle of view and no object. Furthermore, the object area detection unit 102 determines whether a moving object area exists in the view sharing area.
Step S200:
The load estimation unit 103 estimates the processing load in the object recognition unit 105 from the moving object region. The recognition area limiting unit 104 uses the estimated processing load information to mediate which processing load is higher between the image processing apparatuses 100 and 200. The recognition area limiting unit of the image processing apparatus with a high processing load limits the area of the object recognition process performed in step S300.
Step S300:
The object recognizing unit 105 recognizes whether a desired object is shown in the area to be recognized determined in step S200.
As described above, the image processing apparatus 100 repeats the processing from step S100 to step S300 for each input image input by the image acquisition unit 101. Similarly, the image processing apparatus 200 repeats the processing from step S100 to step S300 for each input image acquired by the image acquisition unit 201.

The detailed flow of step S100 will be described below using FIG. FIG. 4 is a flowchart illustrating an example of detailed processing in step S100.
Step S101:
First, the image acquisition unit 101 acquires an image captured by the imaging device 110 such as a video camera. The acquired image is, for example, two-dimensional array data composed of 8-bit pixels, and is composed of R, G, B, and three surfaces. The image acquisition unit 101 may convert RGB data into luminance data and use the luminance image for subsequent processing. In the present embodiment, it is assumed that a moving image is handled, and the processing is performed in units of frames. Therefore, the image obtained here is a still image for each frame.
Step S102:
The object region detection unit 102 determines whether or not a region in the image acquired by the image acquisition unit 101 is set as a shared region between imaging devices provided with adjacent fields of view.
If the setting has been completed, the process branches to S104.
If the setting of the shared area is incomplete, the process branches to S103 to set the shared area.
Step S103:
The object area detection unit 102 sets an area that shares a field of view with an adjacent imaging device in the image output by the image acquisition unit 101.
Step S104:
The object region detection unit 102 compares the image output from the image acquisition unit 101 with a reference model stored in the storage device 11, and detects and outputs a region in which an object is captured from the image. Here, the reference model is a model representing an image of only a background where no object is shown. Hereinafter, details of the process will be described with reference to the drawings.

FIG. 5 is a diagram illustrating an example of an image acquired by the image acquisition unit 101 when a person is imaged by the imaging device 110. It is assumed that the right side region (R) in the figure is a visual field sharing region. A and B are persons to be recognized.
FIG. 6 is a diagram illustrating an example of a background image serving as a reference model stored in the storage device 11. In the example of FIG. 6, the image is acquired by the image acquisition unit 101 in a state where no object is captured.
The reference model is obtained by dividing the background image of FIG. 6 in predetermined block units and storing image feature amounts for each block, as shown in FIG. FIG. 7 is a diagram illustrating an example of the reference model. As an example, the object region detection unit 102 divides an image into 8 × 8 blocks, and performs DCT (discrete cosine transform) processing using an image of 64 pixels in the block. A part of the DCT coefficient that is the processing result is stored as a feature amount. Similarly, in the image shown in FIG. 5, the object region detection unit 102 divides the block into 8 × 8 pixel blocks, performs DCT processing for each block, and obtains a feature amount. The distance D between the feature amount obtained for each block of the object region detection unit 102 and the feature amount of the corresponding block in the reference model is obtained by the following (Equation 1).
^{D = a t · | X-} Xb | ( Equation 1)
However, X and Xb are feature vectors each constituted by the image feature amount of each block. a is a predetermined discrimination vector for determining the object region. -Is a vector dot product operation. || is an absolute value calculation.
When the distance D obtained by the above (Equation 1) is equal to or greater than a predetermined threshold, the object area detection unit 102 obtains a binary image by setting the block to 1 and setting the block smaller than the predetermined threshold to 0. FIG. 8 shows the results obtained from the images of FIGS. 5 and 6, assuming that 1 block is black and 0 block is white. FIG. 8 is a diagram illustrating an example of an image showing a moving object region extracted by the object region detecting unit 102. In the binary image, a black part indicates a block in which an object exists, and a white part indicates a block only in the background in which no object exists.
Note that the reference model stored in the storage device 11 is updated in a block that is determined from the result of the object region detection unit 102 only as a background where no object exists. For example, this is performed by a weighted average of the feature amount of the image street block acquired by the image acquisition unit 101 and the feature amount of the reference model stored in the storage device 11.

Step S105:
The object area detection unit 102 determines whether or not a block in which an object having a value of 1 exists in the visual field sharing area set in step S103 in the binary image obtained in step S104. If there is a moving object area in the view sharing area, the process branches to step S200. If not, the process branches to step S300 without step S200.
The image processing apparatus 200 performs the same flow. FIG. 9 is a diagram illustrating an example of an image acquired by the image acquisition unit 201 when the image capturing apparatus 210 captures an image at the same time as in FIG. In addition, the left area (S) in the figure is a visual field sharing area. B is a person to be recognized. FIG. 10 is a diagram illustrating an example of results obtained when step S104 is performed in FIG.

Next, the detailed flow of step S200 is demonstrated using FIG. FIG. 11 is a flowchart illustrating an example of detailed processing in step S200.
Step S201:
The load estimation unit 103 acquires the binary image output from the object region detection unit 102.
The load estimation unit 103 determines the area of the moving object region using the acquired binary image. For example, the load estimation unit 103 calculates the area by calculating the total number of blocks having a value of 1 as a moving object region among the total number of blocks in the binary image.
Step S202:
The recognition area limiting unit 104 transmits the area of the moving object area in the image obtained by the load estimation unit 103 in step S200 to the image processing apparatus 200 sharing the visual field via the communication unit 106. Similarly, the recognition area limiting unit 204 of the image processing apparatus 200 transmits area information of the moving object area to the image processing apparatus 100 via the communication unit 206. Here, the moving body region area information transmitted from the communication unit 206 to the communication unit 106 is the moving body region area information calculated by the load estimation unit 203 based on the image acquired by the image acquisition unit 201.

Step S203:
The recognition area limiting unit 104 acquires area information of the moving object area of the image processing apparatus 200 from the communication unit 106. Then, the recognition region limiting unit 104 compares the area information of the moving body region calculated by the load estimating unit 103 with the acquired area information of the moving body region. Here, when the area is obtained by the total number of blocks of the moving object region in the binary image, the recognition region limiting unit 104 may compare the number of blocks.
If it is determined that the area of the moving object region calculated by the load estimation unit 103 is larger, the process branches to step S204. If it is determined that the area of the moving object area acquired from the image processing apparatus 200 is larger, the process branches to step S205.
Step S204:
In step S203, the recognition area limiting unit 104 determines that the area of the moving object area in its own image is larger than the area of the moving object area in the image of the image processing apparatus 200, and performs object recognition processing only for areas other than the shared area. Limit the area to do so.

Step S205:
In step S203, the recognition region limiting unit 104 determines that the area of the moving object region in its own image is smaller than the area of the moving object region in the image of the image processing apparatus 200, and the entire image region (entire range). An area is designated to perform object recognition processing.
8 and 10, the area of the moving object area in the image of the image processing apparatus 100 is larger than the area of the moving object area in the image of the image processing apparatus 200. Therefore, the image processing apparatus 100 sets the object recognition range to an area excluding the view sharing area so that the object recognition process is performed only in areas other than the sharing area. Conversely, the image processing apparatus 200 determines to perform object recognition processing for the entire area including the shared area.

Next, an example of the process of step S300 will be described.
Step S300:
The object recognition unit 105 uses the object dictionary stored in the storage device 11 to check whether or not a desired object is reflected in the image output from the image acquisition unit 101. At this time, in step S200, the area recognized by the object recognition unit 105 is switched to the entire area including the shared area or the area excluding the shared area.
In the present embodiment, a person is assumed as an object to be recognized, and the method proposed in Non-Patent Document 1 can be applied to object recognition. The method proposed in Non-Patent Document 1 is applied to the region determined as the moving object region by the object region detection unit 102 and the region limited by the recognition region limitation unit 104.
By performing the above operation in the image processing apparatus 200 in the same manner at the same time, the image processing apparatus 100 and the image processing apparatus 200 perform arbitration so that the object recognition processing of the object existing in the shared area is performed by a lower processing load. be able to. Therefore, the object recognition processing load can be distributed among the image processing apparatuses.

<Embodiment 2>
FIG. 12 is a diagram illustrating an example of a system configuration and the like of the image processing system according to the second embodiment. In the present embodiment, an image processing system that performs face authentication of a specific person from an image will be described as an example.
The imaging device 310 is a device that captures an image, such as a video camera, and includes an imaging lens and an imaging element that performs photoelectric conversion to generate a pixel signal. Note that an image captured by the imaging device 310 is input to the image acquisition unit 301.
The image acquisition unit 301 acquires an image captured by the imaging device 310.
The object area detection unit 302 extracts a specific image area from the image acquired by the image acquisition unit 301.
In the present embodiment, the object area detection unit 302 extracts a human face area from the image acquired by the image acquisition unit 301.
The load estimation unit 303 estimates a processing load when the object recognition unit 305 performs the recognition process from the image region extracted by the object region detection unit 302.
The recognition area limiting unit 304 transmits the processing load amount estimated by the load estimation unit 303 to the image processing apparatus 400 via the communication unit 306. Conversely, when the processing load amount is transmitted from the image processing apparatus 400, the recognition area limiting unit 304 compares the processing load amount estimated by the load estimation unit 303. Then, the recognition area limiting unit 304 determines which processing load is lighter between the image processing apparatus 300 and the image processing apparatus 400 and returns a process availability response to the image processing apparatus 400. The recognition area limiting unit 304 limits the area on which the object recognition unit 305 performs recognition processing according to the availability response received from the communication unit 306 or the comparison result of itself. It is assumed that whether the transmission operation or the reception operation is mainly performed can be switched automatically or manually.
The object recognition unit 305 recognizes and detects whether or not a desired object is shown in the specific image region extracted by the object region detection unit 302.
In the present embodiment, the object recognition unit 305 performs face authentication to determine whether a specific person is captured in the face area of the person extracted by the object region detection unit 302.
The communication unit 306 transmits and receives the processing load amount and the availability response.

Similarly, the imaging apparatus 410 is an apparatus that captures an image such as a video camera, and includes an imaging lens and an imaging element that performs photoelectric conversion to generate a pixel signal. The captured image is input to the image acquisition unit 401.
The image acquisition unit 401 acquires an image captured by the imaging device 410.
The object area detection unit 402 extracts a specific image area from the image acquired by the image acquisition unit 401.
In the present embodiment, the object region detection unit 402 extracts a human face region from the image acquired by the image acquisition unit 401.
The load estimation unit 403 estimates a processing load when the object recognition unit 405 performs the recognition process from the image region extracted by the object region detection unit 402.
The recognition area limiting unit 404 transmits the processing load amount estimated by the load estimation unit 403 to the image processing apparatus 300 via the communication unit 406. Conversely, when the processing load amount is transmitted from the image processing apparatus 300, the recognition area limiting unit 404 compares the processing load amount estimated by the load estimation unit 403. The recognition area limiting unit 404 determines which processing load is lighter between the image processing apparatus 400 and the image processing apparatus 300, and returns a process availability response to the image processing apparatus 300. The recognition area limiting unit 404 limits the area in which the object recognition unit 405 performs recognition processing according to the availability response received from the communication unit 406 or the comparison result of itself. It is assumed that whether the transmission operation or the reception operation is mainly performed can be switched automatically or manually.
The object recognition unit 405 recognizes and detects whether or not a desired object is shown in the specific image area extracted by the object area detection unit 402.
In the present embodiment, the object recognizing unit 405 performs face authentication to determine whether a specific person appears in the face area of the person extracted by the object area detecting unit 402.
The communication unit 406 transmits / receives a processing load and availability response.

FIG. 13 is a flowchart illustrating an outline of the second embodiment. The following will be described in order.
Step S400:
First, the object area detection unit 302 extracts a human face area from the image acquired by the image acquisition unit 301. Furthermore, the object area detection unit 302 determines whether a face area exists in the shared area.
Step S500:
The load estimation unit 303 counts the total number of human face areas and estimates the processing load in the object recognition unit 305. The load estimation unit 303 uses the estimated processing load information to mediate which processing load is higher between the image processing apparatuses 300 and 400. The recognition area limiting unit of the image processing apparatus with a high processing load limits the area of the object recognition process performed in step S600.
Step S600:
The object recognizing unit 305 performs face authentication to determine whether a specific person is shown in the area to be recognized in step S500.
As described above, the image processing apparatus 300 repeats the processing from step S400 to step S600 for each input image input by the image acquisition unit 301. Similarly, the image processing apparatus 400 repeats the processing from step S400 to step S600 for each input image acquired by the image acquisition unit 401.

Hereinafter, the detailed flow of step S400 will be described with reference to FIG. FIG. 14 is a flowchart illustrating an example of detailed processing in step S400.
Step S401:
First, the image acquisition unit 301 acquires an image captured by an imaging device 310 such as a video camera. The acquired image is, for example, two-dimensional array data composed of 8-bit pixels, and is composed of R, G, B, and three surfaces. The image acquisition unit 301 may convert the RGB data into luminance data and use the luminance image for subsequent processing. In the present embodiment, it is assumed that a moving image is handled, and the processing is performed in units of frames. Therefore, the image obtained here is a still image for each frame.
Step S402:
The object region detection unit 302 determines whether or not a region in the image acquired by the image acquisition unit 301 is a shared region between the imaging devices provided with adjacent fields of view.
If the setting is completed, the process branches to S404.
If the setting of the shared area is incomplete, the process branches to S403 to set the shared area.

Step S403:
In the image output by the image acquisition unit 301, an area sharing a field of view with an adjacent imaging device is set.
Step S404:
The object area detection unit 302 detects the face area of the person from the image by performing the same processing as in steps S104 and S300 of the first embodiment on the image output from the image acquisition unit 301.
Step S405:
The object area detection unit 302 determines whether or not the face area block obtained in step S404 exists in the shared area set in step S403. If an object area exists in the shared area, the process branches to step S500. If not, the process branches to step S600 without step S500.

Next, the detailed flow of step S500 is demonstrated using FIG. 15, FIG. FIG. 15 is a flowchart illustrating an example of processing of the image processing apparatus 300. FIG. 16 is a flowchart illustrating an example of processing of the image processing apparatus 400. First, FIG. 15 will be described.
Step S501:
The load estimation unit 303 acquires an image from which the face area output from the object area detection unit 302 is extracted.
The load estimation unit 303 performs contour extraction using the acquired image and determines the total number of human faces in the image.
Step S502:
The recognition area limiting unit 304 processes the image of the imaging device 410 that shares the field of view with the imaging device 310 via the communication unit 306, using the total number of faces in the image obtained by the load estimation unit 303 in step S400. To device 400.
Step S503:
The communication unit 306 receives an overlap area processing availability response from the image processing apparatus 400 and sends it to the recognition area limiting unit 304.
Step S504:
The recognition area limiting unit 304 acquires from the communication unit 306 a response indicating whether the shared area of the image processing apparatus 400 can be recognized. If the response is acceptable, the process branches to S505. If the response is negative, the process branches to S506.
Step S505:
In step S504, the recognition area limiting unit 304 determines that the image processing apparatus 400 performs face authentication processing in the shared area, and the object recognition range is an area excluding the visual field shared area so that the object recognition process is performed only in areas other than the shared area. Set to.
Step S506:
In step S503, the recognition area limiting unit 304 determines that the image processing apparatus 400 does not perform face authentication processing in the shared area, and designates an area so that object recognition processing is performed in the entire image area (entire range).

The detailed flow of the image processing apparatus 400 has been described above. Next, FIG. 16 will be described. The same numbered flow as in FIG. 15 performs the same processing.
Step S507:
The recognition area limiting unit 404 of the image processing apparatus 400 receives the total number information of the face areas of the image processing apparatus 300 from the communication unit 406.
Step S508:
The recognition area limiting unit 404 compares the received face area total number information of the image processing apparatus 300 with the face area total number information counted by the load estimation unit 403, and determines which is larger. If the total number of face areas in the image of the image processing apparatus 400 is larger, the process branches to step S509. If the total number of face areas in the image of the image processing apparatus 300 is larger, the process branches to step S510. In the case of the same number, it may branch to either. In the present embodiment, it is assumed that the process branches to step S510.
Step S509:
The recognition area limiting unit 404 transmits that the face authentication process for the shared area is not performed to the image processing apparatus 300 via the communication unit 406.
Step S510:
The recognition area limiting unit 404 transmits to the image processing apparatus 300 that the face authentication process for the shared area is performed via the communication unit 406.

Next, an example of detailed processing in step S600 will be described.
Step S600:
The object recognition unit 305 uses the object dictionary stored in the storage device 11 to check whether a desired object is reflected in the image output from the image acquisition unit 301. At this time, in step S500, the area recognized by the object recognition unit 305 is switched between the entire area including the shared area or the area excluding the shared area. Note that in the case of a system that branches from step S400 to step S500 and branches to step S600, the region recognized by the object recognition unit 305 is set to all regions including the shared region.
In the present embodiment, a face of a specific person is assumed as an object to be recognized, and a method proposed in Reference Document 1 described later can be applied to person authentication.
First, the face area is extracted in step S404. The object recognition unit 305 extracts feature points such as eyes and nose from the extracted face area. The object recognizing unit 305 cuts out the face area to a normalized size using the extracted feature points as a reference, and extracts grayscale information and the like as feature amounts. Individual feature amounts extracted from the face image of the person to be recognized are stored in the storage device 11 in advance. The object recognizing unit 305 can identify the face of an individual by calculating the similarity between the individual feature amount stored in the storage device 11 and the feature amount extracted from the image.
With the above operation, the image processing apparatus 300 and the image processing apparatus 400 can perform arbitration so that the object recognition processing of the object existing in the shared area is performed by a lower processing load.
Reference 1 Yamaguchi. Fukui, “Face recognition system“ Smartface ”” robust to changes in face orientation and facial expression, IEICE Transactions, D-II, Vol. J84-D-II, no. 6, pp. 1045-1052, 2001-6

<Other embodiments>
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program. It is a process to be executed.

  As mentioned above, according to each embodiment mentioned above, in the device which picturizes a wide area using a plurality of imaging devices, it becomes possible to reduce processing load of an imaging device with high recognition processing load. That is, it is possible to reduce the frequency with which the load of recognition processing is high and the frames are dropped because the processing is not in time. That is, recognition processing can be reduced.

  The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to such specific embodiments, and various modifications can be made within the scope of the gist of the present invention described in the claims.・ Change is possible.

DESCRIPTION OF SYMBOLS 10 Control apparatus, 100 Image processing apparatus, 110 Imaging device

Claims (8)

Second image processing that performs image processing of an image captured by the first imaging means and performs image processing of an image captured by the second imaging means that shares a part of the imaging range with the first imaging means An image processing apparatus that communicates with an apparatus,
Obtaining means for obtaining an image captured by the first imaging means;
Detection means for detecting an object region from the image acquired by the acquisition means;
If the object area detected by the detection means is in a shared area between the first imaging means and the second imaging means in the acquired image, the object recognition processing of the shared area is performed as the image processing. Determination means for determining which of the apparatus and the second image processing apparatus to perform;
When it is determined by the image processing apparatus that the determination means performs the object recognition range, the object recognition range is set to the entire range of the image, and when it is determined by the determination means that the second image processing apparatus performs the object recognition, Setting means for setting a range to an area excluding the shared area from the entire range of the image;
Recognizing means for recognizing whether or not an object is reflected in the object recognition range set by the setting means;
An image processing apparatus. A calculation means for calculating the area of the moving object region in the acquired image;
The determination unit transmits the area calculated by the calculation unit to the second image processing device, and the moving object in the image captured by the second imaging unit calculated by the second image processing device. When the area of the area is received from the second image processing apparatus and the calculated area is larger than the received area, it is determined that the object recognition process of the shared area is performed by the second image processing apparatus. The image processing apparatus according to claim 1, wherein when the calculated area is not larger than the received area, the image processing apparatus determines that the object recognition process of the shared area is performed by the image processing apparatus. A counting means for counting the total number of face regions in the image;
The determination means transmits the total number of the face areas counted by the counting means to the second image processing apparatus, and whether or not the second image processing apparatus can recognize the object in the shared area. If the response is received from the second image processing apparatus and the received availability response indicates that the object recognition of the shared area is possible, the object recognition process of the shared area is changed to the second image processing. 2. The apparatus according to claim 1, wherein the image recognition apparatus determines that the image processing apparatus performs object recognition processing of the shared area when it is determined that the apparatus recognizes that the object is not recognized in the shared area. Image processing apparatus.   The image processing apparatus according to claim 1, wherein the recognition unit recognizes whether a human body is in the object recognition range set by the setting unit.   The image processing apparatus according to claim 1, wherein the recognizing unit recognizes whether a face of a specific person is in the object recognition range set by the setting unit. An image processing apparatus that performs image processing of an image captured by the first imaging means, and a second image processor that performs image processing of an image captured by the second imaging means that partially shares the imaging range with the first imaging means. An image processing system comprising: a second image processing apparatus;
The first image processing apparatus includes:
Obtaining means for obtaining an image captured by the first imaging means;
Detection means for detecting an object region from the image acquired by the acquisition means;
If the object area detected by the detection means is in the shared area between the first imaging means and the second imaging means in the acquired image, the object recognition processing of the shared area is performed in the first area. Determination means for determining which of the image processing apparatus and the second image processing apparatus to perform,
When it is determined by the determination means that the first image processing apparatus is used, the object recognition range is set to the entire range of the image, and when the determination means is determined to be performed by the second image processing apparatus Setting means for setting the object recognition range to an area excluding the shared area from the entire range of the image;
Recognizing means for recognizing whether or not an object is reflected in the object recognition range set by the setting means;
Have
The second image processing apparatus includes:
Obtaining means for obtaining an image captured by the second imaging means;
Detection means for detecting an object region from the image acquired by the acquisition means;
When the object area detected by the detection means is in the shared area between the second imaging means and the first imaging means in the acquired image, the object recognition processing of the shared area is performed in the second area. Determining means for determining which of the image processing apparatus and the first image processing apparatus to perform,
When it is determined by the determination means that the second image processing apparatus is used, the object recognition range is set to the entire range of the image, and when the determination means is determined to be performed by the first image processing apparatus Setting means for setting the object recognition range to an area excluding the shared area from the entire range of the image;
Recognizing means for recognizing whether or not an object is reflected in the object recognition range set by the setting means;
An image processing system. Second image processing that performs image processing of an image captured by the first imaging means and performs image processing of an image captured by the second imaging means that shares a part of the imaging range with the first imaging means An image processing method executed by an image processing apparatus that communicates with an apparatus,
An acquisition step of acquiring an image captured by the first imaging means;
A detection step of detecting an object region from the image acquired in the acquisition step;
When the object area detected in the detection step is in a shared area between the first imaging means and the second imaging means in the acquired image, the object recognition processing of the shared area is performed as the image processing. A determination step of determining which of the apparatus and the second image processing apparatus to perform;
If it is determined in the determination step that the image processing apparatus performs the object recognition range, the object recognition range is set to the entire range of the image. If it is determined in the determination step that the second image processing apparatus performs the object recognition, A setting step for setting a range to an area excluding the shared area from the entire range of the image;
A recognition step of performing object recognition as to whether or not an object is reflected in the object recognition range set in the setting step;
An image processing method including: An image processing apparatus that performs image processing of an image captured by the first imaging means, and a second image processor that performs image processing of an image captured by the second imaging means that partially shares the imaging range with the first imaging means. An image processing method in an image processing system, comprising:
The first image processing device acquires an image captured by the first imaging unit; and
The first image processing apparatus detects an object region from the image acquired in the acquisition step;
When the first image processing apparatus includes the object area detected in the detection step in the shared area of the first imaging means and the second imaging means in the acquired image, the sharing A determination step of determining whether the object recognition processing of the area is performed by the first image processing apparatus or the second image processing apparatus;
When the first image processing device is determined to be performed by the first image processing device in the determination step, the object recognition range is set to the entire range of the image, and the second image is determined in the determination step. When it is determined to be performed by the processing device, a setting step for setting an object recognition range to an area excluding the shared area from the entire range of the image;
A recognition step in which the first image processing apparatus performs object recognition as to whether or not an object is reflected in the object recognition range set in the setting step;
An acquisition step in which the second image processing apparatus acquires an image captured by the second imaging unit;
The second image processing apparatus detects a object region from the image acquired in the acquisition step;
When the second image processing apparatus includes the object region detected in the detection step in a shared region between the second imaging unit and the first imaging unit in the acquired image, the sharing A determination step of determining whether the object recognition processing of the region is performed by the second image processing device or the first image processing device;
When the second image processing device is determined to be performed by the second image processing device in the determination step, the object recognition range is set to the entire range of the image, and the first image is determined in the determination step. When it is determined to be performed by the processing device, a setting step for setting an object recognition range to an area excluding the shared area from the entire range of the image;
A recognition step in which the second image processing apparatus performs object recognition as to whether or not an object is reflected in the object recognition range set in the setting step;
An image processing method including:

Images