JP2018124629A - Image processing device, information processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To more easily increase speed in detection processing of an object.SOLUTION: An image processing device detects an object, based on size information related to the size of the object which becomes a detection target in an area, from the area set to an image, and updates the size information, based on the size of the detected object.SELECTED DRAWING: Figure 6

Description

  The present invention relates to an image processing apparatus, an information processing method, and a program.

Conventionally, when an object such as a human body or a face is detected from an image taken by an imaging device such as a monitoring camera, the object is detected by pattern matching processing. In such a case, the size of the object in the captured image differs depending on the angle of view of the imaging device, the distance between the imaging device and the subject, etc., and therefore various types of object detection processing are performed on the entire image. It was broken. However, it takes a lot of time to detect objects of various sizes on the entire image. Therefore, in order to speed up the processing, the angle of view of the imaging device, the distance between the imaging device and the subject is set, or measurement is performed using the measurement device, thereby detecting from the image taken by the imaging device. A method for limiting the size of an object has been proposed.
For example, Patent Document 1 discloses a method for determining a face size range to be detected according to distance information obtained by a distance sensor.
Patent Document 2 discloses a method for calculating an actual size of an object by measuring a distance to the object photographed by a stereo camera.

JP 2004-318331 A JP 2008-236276 A

However, in the conventional techniques disclosed in Patent Documents 1 and 2, the distance between the imaging device and the subject must be measured by some method. If there is no information on the distance, the size of the object to be detected is reduced. There is a problem that can not be. In addition, it is necessary to set information such as the angle of view of the imaging apparatus and the distance between the imaging apparatus and the subject every time the arrangement position of the imaging apparatus is changed.
The present invention has been made to solve the above-described problems, and an object thereof is to speed up the object detection process more simply.

  An image processing apparatus according to the present invention includes: a detection unit that detects an object based on size information relating to a size of an object to be detected in the region from a region set for the image; and the detection unit detects the object Updating means for updating the size information based on the size of the object.

  According to the present invention, it is possible to speed up the object detection process more easily.

It is a figure which shows an example of the hardware constitutions of an image processing apparatus. It is a figure which shows an example of a function structure of an image processing apparatus. It is a figure explaining an example of the field included in an image. It is a figure explaining an example of the information of detection size. It is a figure explaining an example of the detection process of an object. It is a flowchart which shows an example of a process of an image processing apparatus. It is a figure explaining an example of the process in the case of a view angle change. It is a figure explaining an example of the image | photographed image. It is a figure explaining an example of the division | segmentation of an image. It is a figure explaining an example of a display of the size of the object of detection set up for every field.

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

<Embodiment 1>
Assume that the image processing apparatus 100 according to the present embodiment is an imaging apparatus such as a surveillance camera having an image processing function.
FIG. 1 is a diagram illustrating an example of a hardware configuration of the image processing apparatus 100. The image processing apparatus 100 includes a CPU 101, a main storage device 102, an auxiliary storage device 103, a photographing unit 104, a network I / F 105, and an input / output I / F 106. Each element is connected to be communicable with each other via a system bus 107.
The CPU 101 is a central processing unit that controls the image processing apparatus 100. The main storage device 102 is a storage device such as a random access memory (RAM) that functions as a work area for the CPU 101 or a temporary storage area for data. The auxiliary storage device 103 is a storage device such as a read only memory (ROM) or a hard disk drive (HDD) that stores various setting information, various programs, information on images captured by the image capturing unit 104, and the like. Further, the auxiliary storage device 103 may be a solid state drive (SSD) or the like.

The imaging unit 104 is an imaging unit that captures an image via an imaging element such as a CCD, a lens, or the like. The network I / F 105 is an interface used for communication with an external device via a network. The input / output I / F 106 is an interface used for inputting / outputting information to / from an external device such as an external storage device.
In the present embodiment, the CPU 101 performs processing for detecting an object from an image photographed by the photographing unit 104. In the present embodiment, the object to be detected is a human body, but may be other objects such as a face, a vehicle, a cargo, and an animal.
The CPU 101 executes processing based on a program stored in the auxiliary storage device 103, thereby realizing the functions of the image processing apparatus 100 described later with reference to FIG. 2, the flowchart processing described with reference to FIG.

FIG. 2 is a diagram illustrating an example of a functional configuration of the image processing apparatus 100. The function of the image processing apparatus 100 will be described with reference to FIG.
The image processing apparatus 100 includes an image acquisition unit 201, a detection unit 202, an output unit 203, a size input unit 204, and a pattern management unit 205. The image acquisition unit 201 acquires an image (moving image or still image) acquired from the imaging unit 104 or from the outside, and transmits the acquired image to the detection unit 202. The image acquisition unit 201 may acquire an image from an external imaging device or a server device that supplies an image. Further, the image acquisition unit 201 may acquire an image stored in the auxiliary storage device 103 in the image processing apparatus 100. The image acquired by the image acquisition unit 201 is an example of a detection target image input to the image processing apparatus 100.

When the moving image is acquired as an image, the image acquiring unit 201 sequentially transmits each frame image constituting the acquired moving image to the detecting unit 202. Then, the detection unit 202 performs an object detection process for each transmitted frame. When acquiring a still image as an image, the image acquisition unit 201 transmits the acquired still image to the detection unit 202. Hereinafter, a case where one frame in a still image or a moving image is transmitted from the image acquisition unit 201 to the detection unit 202 will be described.
The detection unit 202 includes a detection control unit 210, a size setting unit 211, a selection unit 212, an area setting unit 213, a determination unit 214, and a size management unit 215.
The detection control unit 210 mainly controls each element in the object detection process.

FIG. 3 is a diagram illustrating an example of a region included in an image. In the example of FIG. 3, a state in which the image transmitted from the image acquisition unit 201 is divided into a plurality of regions having the same size is shown. The size setting unit 211 sets a plurality of areas to be subjected to object detection processing in the image transmitted from the image acquisition unit 201 by a preset method. For example, the size setting unit 211 sets each region obtained by dividing the image transmitted from the image acquisition unit 201 into N equal parts in the vertical and horizontal directions as the target region for object detection processing. The object detection target region set in the image transmitted from the image acquisition unit 201 is an example of a region set for the image.
The size setting unit 211 sets and sets the size range of the object to be detected in each region set in the image transmitted from the image acquisition unit 201 (for example, each region in FIG. 3). The size management unit 215 is instructed to store the size information. In the following, the size of an object to be detected in each area is referred to as a detection size. The detection unit 202 performs processing for detecting an object having a size indicated by the detection size information for each region in the image. The size management unit 215 stores and manages information on the detected size set by the size setting unit 211 in the auxiliary storage device 103 or the like.

FIG. 4 is a diagram illustrating an example of detection size information managed by the size management unit 215. The table in FIG. 4 includes an area item indicating each area in the image and a detection size item. In the example of FIG. 4, the detection size information is information of a combination of the minimum size and the maximum size, and an object having a size that is equal to or larger than the size indicated by the minimum size and smaller than the size indicated by the maximum size is to be detected. Indicates. In the example of FIG. 4, the minimum size and the maximum size indicate the length of one side of the rectangular area including the detected object.
The size input unit 204 accepts designation of a detection size for each area in the image based on, for example, an operation by the user via the operation unit of the image processing apparatus 100. Also, the size input unit 204 receives, for example, detection size information for each area in the image from an external information processing apparatus via the network I / F 105, thereby detecting the detection size for each area in the image. It is also possible to accept the designation.
The detection size information managed by the size management unit 215 includes detection size information in the example of FIG. 4, but other information such as the number of detection processes and the average size of detected objects are further included. It may be included.

Next, an outline of object detection processing by the detection unit 202 will be described. In the present embodiment, the detection unit 202 performs an object detection process as follows. First, the detection unit 202 arranges a collation pattern in the upper left corner of an image to be subjected to object detection processing. In this embodiment, the position of the collation pattern in the image to be detected is set as the pixel position of the upper left corner of the collation pattern in the image. That is, the detection unit 202 arranges the collation pattern so that the upper left corner of the target image on which the object detection process is performed overlaps the upper left corner of the collation pattern. A collation pattern is an image used for pattern matching in object detection processing. Pattern matching is a process for determining whether or not a specific pattern matches or approximates a comparison target pattern.
As the pattern matching process, for example, the difference between the pixel of the matching pattern and the pixel of the image to be detected that overlaps the pixel of the matching pattern is totalized for all the pixels of the matching pattern, There is a process for determining a threshold value. In this case, for example, if the summed difference value is equal to or smaller than the set threshold value, the detection unit 202 is an object that indicates a portion that overlaps with the matching pattern in the detection target image as the matching pattern. judge. Further, for example, if the summed difference value is larger than the set threshold value, the detection unit 202 determines that a portion overlapping with the matching pattern in the detection target image is not the object indicated by the matching pattern.
As pattern matching processing, for example, the following processing is also available. In other words, this is a process of extracting a feature amount from each of the collation pattern and the detection target image overlapping the collation pattern, and determining a threshold value of the difference value of the extracted feature amount. The extracted feature quantities include HAAR-LIKE feature quantities, HOG feature quantities, SHIFT feature quantities, and the like.

Hereinafter, the object detection process in the present embodiment will be described more specifically.
The detection control unit 210 transmits the image transmitted from the image acquisition unit 201 (hereinafter, the first image) to the selection unit 212.
The selection unit 212 generates a plurality of images having different resolutions by reducing the first image in order to detect objects of different sizes from the first image transmitted from the detection control unit 210.
Here, the significance of generating a plurality of images obtained by reducing the first image will be described. In the present embodiment, the detection unit 202 performs pattern matching on the overlapping portion between the matching pattern and the first image that is the detection target image, so that the portion overlapping the matching pattern is an object indicated by the matching pattern. It is determined whether or not there is. However, the size of the object that can be detected by pattern matching with the matching pattern is only the same size as the matching pattern. In the present embodiment, the detection unit 202 needs to detect objects of various sizes from the first image. Therefore, the detection unit 202 can detect objects of various sizes by generating an image obtained by reducing the first image and performing pattern matching with the same matching pattern on the generated image. For example, when the matching pattern is a 20-pixel square rectangle, the detection unit 202 can detect a 20-pixel object from the first image and can detect a 40-pixel object from an image obtained by reducing the first image by half. It becomes like this. In other words, the selection unit 212 generates a reduced image of the first image according to the size of the object to be detected.

The selection unit 212 generates a reduced image of the first image. At this time, the number of reduced images is determined so that an object can be detected within the detection size range set by the size setting unit 211, and a detection target image group is obtained. For example, it is assumed that the detection size for each region existing in the upper half of the first image has a minimum size of 20 pixels and a maximum size of 130 pixels. Further, it is assumed that the detection size for each region existing in the lower half of the first image is 100 pixels for the minimum size and 150 pixels for the maximum size. In that case, the selection unit 212 generates a reduced image of the first image in multiple stages so that, for example, an object having a size from 20 pixels to 150 pixels can be detected. For example, when the selection unit 212 wants to detect an object smaller than the size of the matching pattern, the selection unit 212 may generate an enlarged image of the first image. In this case, the detection unit 202 performs an object detection process using the matching pattern on the enlarged image of the first image.
FIG. 5 is a diagram illustrating an example of the first image and a reduced image of the first image. The image 411 is a first image, and the images 412 and 413 are reduced images of the first image. A collation pattern 401 is a collation pattern used for object detection processing in the present embodiment, and is an image showing a human body that is an object to be detected.

Next, the detection unit 202 performs object detection processing using the matching pattern 401 on the reduced image of the first image generated by the first image and selection unit 212. The selection unit 212 selects one image to be detected from the first image and the reduced image one by one, and transmits the selected image to the region setting unit 213. Here, processing when the selection unit 212 selects the image 411 will be described. For example, the determination unit 214 acquires information on the verification pattern 401 from the auxiliary storage device 103 and the like via the pattern management unit 205 and transmits the information to the region setting unit 213.
The area setting unit 213 arranges the collation pattern 401 at a preset initial position, and moves the collation pattern 401 by a preset movement method. The area setting unit 213 arranges the matching pattern 401 at the upper left corner of the image 411. The area setting unit 213 sequentially moves the collation pattern 401 from the left side to the right side, from top to bottom. For example, the region setting unit 213 moves the collation pattern 401 from the left side to the right side by set pixels (for example, one pixel at a time, five pixels, etc.), and the right corner of the collation pattern 401 is the right corner of the image 411. Move until it overlaps. Next, the region setting unit 213 moves the matching pattern 401 to the left corner, and moves it downward by a set pixel (for example, 1 pixel, 5 pixels, etc.). Then, the area setting unit 213 again moves the matching pattern 401 from the left side to the right side by the set pixels until the right corner of the matching pattern 401 overlaps the right corner of the image 411. The region setting unit 213 repeats the above processing until the lower right corner of the matching pattern 401 overlaps with the lower right corner of the image 411.

In the present embodiment, a plurality of areas are set in the first image, and the size of the object to be detected is set for each area. Therefore, each time the collation pattern 401 is moved by the area setting unit 213, the determination unit 214 performs the following processing. That is, the determination unit 214 specifies to which region of the image 411 the position of the matching pattern 401 (pixel position at the upper left corner) belongs. Then, the determination unit 214 determines whether or not the range indicated by the detection size corresponding to the identified area includes the size of the object corresponding to the image (image 411) currently being detected. If the determination unit 214 determines that it is included, the object detection processing is performed by performing pattern matching between the matching pattern 401 at the current position and the region of the image 411 that overlaps the matching pattern 401 at the current position. Do. If the determination unit 214 determines that the pattern is not included, the determination unit 214 does not perform object detection processing by pattern matching between the matching pattern 401 at the current position and the region of the image 411 that overlaps the matching pattern 401 at the current position. As a result, the image processing apparatus 100 can reduce the processing load and speed up the processing without performing detection processing of an object having a size other than the detection size.
When the determination unit 214 detects an object, the detection control unit 210 outputs the position information of the matching pattern 401 and the size information of the detected object to the size setting unit 211. The size setting unit 211 updates the detection size of the area including the collation pattern 401 based on the position information of the collation pattern 401 output from the detection control unit 210 and the size information of the detected object. In the present embodiment, it is assumed that the matching pattern 401 is included in the region where the matching pattern 401 is located (the region where the upper left corner pixel of the matching pattern 401 exists).
Similarly, the detection unit 202 performs object detection processing on the reduced images of the first image such as the image 412 and the image 413 that are reduced images of the first image. The detection unit 202 updates the detection size after converting the position information and size information indicated by the object detection result for the reduced image of the first image into the position information and size information in the first image.

Next, detection size update processing by the size setting unit 211 will be described. As shown in FIG. 4, a case will be described in which a new object is detected by the determination unit 214 while the size management unit 215 manages information on the detection size for each region. The size setting unit 211 receives the position information and size information of the object detected from the determination unit 214. Then, the size setting unit 211 updates the detection size based on the following formulas 1 and 2.
HBSmin = (HBS_new + HBSmin_old) * γ / 2 (Formula 1)
HBSmax = (HBS_new + HBSmax_old) * γ / 2 (Formula 2)
In Equations 1 and 2, HBSmin is the minimum size in the detection size after update. HBSmax is the maximum size in the detection size after update. HBS_new is the size of the detected object. HBSmin_old is the minimum size in the detection size before update. HBSmax_old is the maximum size in the detection size before update. γ is a size coefficient for adjusting the detection size after update, and takes an arbitrary value. The size setting unit 211 may appropriately change the value of γ.
The size setting unit 211 can reflect the result of the size of the object detected in the past as statistical information on the detection size by updating the detection size using Expressions 1 and 2. After the detection size is updated, the detection unit 202 detects an object having a useless size by detecting an object in the updated detection size range (HBSmin to HBSmax) in consideration of past detection results. Processing can be omitted.

The processing of the detection unit 202 when an image that is a still image is transmitted from the image acquisition unit 201 has been described above. When an image that is a moving image is transmitted from the image acquisition unit 201, the detection unit 202 repeatedly performs the processing described in the present embodiment for each frame included in the moving image.
The output unit 203 outputs information indicating the processing result by the detection unit 202. The output unit 203 outputs information indicating the processing result to, for example, a display unit of the image processing apparatus 100 or a display unit of an external apparatus connected to the image processing apparatus 100 via wireless or wired communication. Further, the output unit 203 may output to a display device such as an external display connected to the image processing apparatus 100 via, for example, wireless or wired connection. The output unit 203 may output information indicating the processing result by the detection unit 202 as it is, or may output information indicating the processing result by the detection unit 202 as visible graphics. The information indicating the processing result by the detection unit 202 is, for example, information indicating the position and size of the detected object, information indicating the detection size of each area updated by the size setting unit 211, and the like.
When the detection unit 202 repeatedly performs these processes to detect objects of various sizes, the process ends. For example, if the detection unit 202 continues the process for a specified period, the process may be completed even if the entire process is not completed, or the number of detected human bodies and the number of times the human body detection process is executed. The process may be terminated when becomes equal to or greater than the set threshold.

FIG. 6 is a flowchart illustrating an example of processing of the image processing apparatus 100. Details of the processing of the present embodiment of the image processing apparatus 100 will be described with reference to FIG.
In step S <b> 501, the size input unit 204 initializes the detection size for each area set in the detection target image via the size setting unit 211. The size input unit 204 accepts designation of the detection size for each region in the image that is the target of detection processing based on, for example, an operation by the user via the operation unit of the image processing apparatus 100. Then, the size input unit 204 instructs the size setting unit 211 to set the detected size indicated by the received designation as the initial value of the detected size of each area. In response to the instruction, the size setting unit 211 stores the detected size value indicated by the designation received by the size input unit 204 in the auxiliary storage device 103 or the like as the initial value of the detected size of each area.
In step S <b> 502, the image acquisition unit 201 acquires an image supplied from the outside. In the example of FIG. 5, it is assumed that the image acquisition unit 201 has acquired a still image. Then, the image acquisition unit 201 transmits the acquired image to the detection unit 202. Then, the detection unit 202 starts object detection processing on the image transmitted from the image acquisition unit 201.
In step S503, the detection control unit 210 determines whether the processing for detecting an object from the image acquired in step S502 is completed. If the detection control unit 210 determines that the process of detecting an object from the image acquired in S502 is completed, the detection control unit 210 proceeds to the process of S507. If the detection control unit 210 determines that the process of detecting an object from the image acquired in S502 has not been completed, the process proceeds to S504.

In step S504, the determination unit 214 detects an object from the image acquired in step S502. The process of S504 is the same process as the process described in FIG. That is, the determination unit 214 selects any one of the image acquired in S502 or the reduced image of the image. Then, the determination unit 214 detects an object of the current detection target size by performing pattern matching using the matching pattern from the selected image.
In step S505, the detection control unit 210 determines whether an object is detected in the detection process in step S504. If the detection control unit 210 determines that the object has been detected in the detection process in S504, the detection control unit 210 transmits the position information and size information of the object detected in S504 to the size setting unit 211, and proceeds to the process of S506. If the detection control unit 210 determines that the object is not detected in the detection process in S504, the process proceeds to S503.
In step S506, the size setting unit 211 identifies in which region the object is detected from the position information of the detected object transmitted in step S505. Then, the size setting unit 211 updates the detected size information corresponding to the identified area based on the size information of the detected object transmitted in S505.
In S507, the detection control unit 210 outputs, for example, information indicating the detection result of the object for the image acquired in S502 to an external output unit such as an external display. The external output unit outputs and outputs the output information.

As described above, in the present embodiment, the image processing apparatus 100 sets the initial value of the detection size information indicating the size of the object to be detected for each set area in the image that is the object of the object detection process. . Then, the image processing apparatus 100 detects an object having a size within a range indicated by the corresponding detection size for each set region in the image that is a target of the object detection processing, and each time the detection is performed, information on the detection size is detected. It was decided to update.
Through such processing, the image processing apparatus 100 can update the size of the object to be detected to an appropriate value for each region without using information on the distance between the image processing apparatus 100 and the subject. In other words, the image processing apparatus 100 can update the detection size information of the set area each time an object is detected, and omit the execution of processing for detecting an object having a useless size. Further, the image processing apparatus 100 does not need to set information such as the angle of view of the imaging apparatus and the distance between the imaging apparatus and the subject each time the arrangement position of the imaging apparatus is changed. Thereby, the image processing apparatus 100 can speed up the object detection process more easily.

In the present embodiment, the image processing apparatus is an imaging apparatus such as a surveillance camera having an image processing function. However, the image processing apparatus 100 may be an information processing apparatus such as a personal computer or a server apparatus that acquires an image captured by an external imaging apparatus and performs image processing on the acquired image.
In the present embodiment, the detection size information is the minimum size information and the maximum size information of the object to be detected in the corresponding region. However, the detection size information may be, for example, information on the median value of the size range of the object to be detected and information on the width of the range.
In this embodiment, the position of the collation pattern in the image to be detected is set as the pixel position of the upper left corner of the collation pattern in the image. However, the position of the collation pattern in the image to be detected may be the position of another pixel such as the center pixel of the collation pattern in the image.
In the present embodiment, the detection unit 202 performs object detection processing by performing pattern matching on the detection target image and its reduced image using a matching pattern. However, the detection unit 202 performs object detection processing by performing pattern matching on the detection target image using a matching pattern and a matching pattern enlarged or reduced to the size of the detection size of each region. It may be done.

In the present embodiment, the size setting unit 211 sets the detection size for each of a plurality of areas into which the image acquired by the image acquisition unit 201 is divided. However, the size setting unit 211 sets detection target areas for the set number (for example, one, two, etc.) in the image acquired by the image acquisition unit 201, and detects the detection size for that area. It is good also as setting. For example, the size setting unit 211 may set the entire image acquired by the image acquisition unit 201 as one region and set the detection size for the region. In addition, the size setting unit 211 sets, for example, one rectangular area having a set size centered on the center of the image acquired by the image acquisition unit 201, and sets a detection size for the area. Also good. In this case, the detection unit 202 does not perform the object detection process for portions other than the rectangular area.
In addition, the size setting unit 211 may accept designation of a region to be detected based on an operation via the operation unit of the image processing apparatus 100 or based on region designation information from an external device. Then, the size setting unit 211 may set an area indicated by the received designation as an area to be detected. By doing so, the image processing apparatus 100 performs the object detection process only on the region desired by the user and does not perform the detection process on other portions, and can reduce the burden of unnecessary processing.
For example, in the image acquired by the image acquisition unit 201, a human body that is an object to be detected cannot be captured for a portion where only a wall can be captured. Therefore, the image processing apparatus 100 may not perform the object detection process by not including such a portion in the region where the detection size is set. Thereby, the image processing apparatus 100 can prevent an increase in unnecessary processing burden.

<Embodiment 2>
In the present embodiment, the processing of the image processing apparatus 100 when the angle of view of an imaging unit (an image processing apparatus 100 having the imaging unit 104, an external imaging apparatus, or the like) that captures an image that is a target of object detection processing is changed. Will be explained. The angle of view of the imaging means is a range photographed by the imaging means.
Hereinafter, differences from the first embodiment will be described.
The hardware configuration and functional configuration of the image processing apparatus 100 of the present embodiment are the same as those of the first embodiment.

In the present embodiment, the image processing apparatus 100 is assumed to be an imaging apparatus such as a surveillance camera having an image processing function, as in the first embodiment. A case where pan / tilt / zoom movement occurs in the image processing apparatus 100 and the angle of view of the image processing apparatus 100 is changed will be described. When the angle of view changes, the size of the object that can be photographed in each area on the image acquired by the image acquisition unit 201 is also changed.
Therefore, when the angle of view of the image processing apparatus 100 is changed, the size setting unit 211 detects the detected size managed by the size management unit 215 based on the change of the angle of view (pan / tilt / zoom movement amount) of the image processing apparatus 100. Update information. For example, when the angle of view is changed to double zoom, the size setting unit 211 updates the minimum size and the maximum size of the detection size of each region to double.
After updating the detection size information, the image processing apparatus 100 detects an object based on the updated detection size information, and changes the angle of view (pan / tilt / zoom movement) of the image processing apparatus 100. The same process is repeated each time.
With the above processing, the image processing apparatus 100 can appropriately update the detection size of each area in the image to be detected when the angle of view of the image processing apparatus 100 is changed.

In addition, when the angle of view of the image processing apparatus 100 is set in advance as a plurality of preset positions, the size management unit 215 manages information on the detected size for each region separately for each preset position. It is good.
FIG. 7 is a diagram for explaining an example of processing for changing the angle of view. The example of FIG. 7 shows a state in which the angle of view at the preset position A, the angle of view at the preset position B, and the angle of view at the preset position C are set in advance for the image processing apparatus 100. The size management unit 215 stores and manages, in the auxiliary storage device 103, detection size lists 611 to 613 that list detection size information of each area corresponding to each preset position. The detection unit 202 specifies whether the current angle of view of the image processing apparatus 100 is the angle of view at the preset position A, the angle of view at the preset position B, or the angle of view at the preset position C. Then, the detection unit 202 detects an object using information on the detected size corresponding to the specified angle of view, and updates the information on the detected size using the result of the detection process. When the angle of view of the image processing apparatus 100 is switched, the detection unit 202 switches detection size information used for the detection process. For example, when switching from the angle of view at the preset position A to the angle of view at the preset position B occurs, the detection unit 202 switches the detection size information to be used from the detection size list 611 to the detection size list 612. Then, the detection unit 202 detects an object using the switched detection size information, and updates the detection size information using the detection result.
With the above processing, the image processing apparatus 100 changes the detection size for the original angle of view that has been updated before, for example, when the angle of view of the image processing apparatus 100 is once changed and then returned to the original angle of view. Can be used. Thereby, the image processing apparatus 100 does not need to update the detection size from the initial value again for the original angle of view, and can reduce the burden of unnecessary processing.

<Embodiment 3>
In the first embodiment, the size setting unit 211 performs the initial setting of the detection size for each of a plurality of regions having the same size set in the detection processing target image. However, in the present embodiment, the size setting unit 211 sets an area to be subjected to object detection processing for an image to be subjected to object detection processing in consideration of how the object appears in the image. To do. The process of this embodiment will be described below.
FIG. 8A is a diagram illustrating an example of a situation in which shooting is performed by the image processing apparatus 100. In the example of FIG. 8A, the image processing apparatus 100 photographs a plurality of persons including persons 701 and 702 existing in the passage via the photographing unit 104. FIG. 8B is an image taken by the image processing apparatus 100 in the situation of FIG.
In the example of FIG. 8A, the image processing apparatus 100 is installed at an angle that looks down the passage obliquely downward. In such an environment, the image captured by the image processing apparatus 100 is an image in which the lower side is a portion that is relatively close to the image processing apparatus 100 and the upper side is a portion that is relatively far from the image processing apparatus 100. The resulting image. In the image of FIG. 8B, the person 701 closest to the image processing apparatus 100 is photographed larger than the other person, and the person 702 farthest from the image processing apparatus 100 is photographed smaller than the other person. Accordingly, the closer to the image processing apparatus 100, the larger the variation in the size of the person in the image due to individual differences such as the height difference of the person. That is, the size of the detected object is predicted to be larger in the lower part of the image than in the upper part of the image.
Based on such predictions, the size setting unit 211 sets a region to be subjected to detection processing as follows for an image shot so as to look down at a plurality of persons. That is, the size setting unit 211 divides an image as shown in FIG. 9 and sets each divided area as an area to be detected. In the example of FIG. 9, each area set in the area 802 below the image is smaller than each area set in the area 801 above the image. The detection unit 202 can update the detection size information for each finer area of the area 802 and can detect the object with respect to the area 801 with higher accuracy.

  With the above processing, the image processing apparatus 100 sets the detection target area in consideration of the appearance of the object in the image, and updates the detection size information for each set area, thereby improving the object accuracy. Can be detected.

<Embodiment 4>
In the present embodiment, a description will be given of a process in which the output unit 203 outputs detection size information as shown in FIG. 4 managed by the size management unit 215.
The output unit 203 outputs the detection size information by displaying it on the display unit of the image processing apparatus 100, an external display, or the like in real time or at a set timing so that the user can check it. Can do.
FIG. 10A is a diagram illustrating an example of a display result of the detection size. The output unit 203 obtains the average size of the objects to be detected corresponding to the area from the information on the detected size for each area. Then, the output unit 203 enlarges / reduces the object type icon according to the obtained average size, and displays it at the location of each region in the image. In this way, the image processing apparatus 100 displays the approximate size of the detection size of each region in association with each region, so that the user can determine the size of the object to be detected in each region. Can be visually grasped.
The detection size for each region is roughly determined by the angle of view. Therefore, it is predicted that the detection size of the adjacent region takes a closer value than the detection size of the distant region. Therefore, when there is an icon 902 indicating a size that is out of the set threshold value or more compared to the surrounding area as in the range 901 of FIG. 10B, the detected size is detected in the area corresponding to the icon 902. You can see visually that the update is not going well. As a result, the image processing apparatus 100 can prompt the user to correct information on the detection size of the area that is not well updated, such as the detection size corresponding to the icon 902.

<Other embodiments>
The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in the computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

For example, a part or all of the functional configuration of the image processing apparatus 100 described above may be implemented in the image processing apparatus 100 as hardware.
As mentioned above, although preferable embodiment of this invention was explained in full detail, this invention is not limited to the specific embodiment which concerns. You may combine each embodiment mentioned above arbitrarily.

100 Image processing apparatus 101 CPU
202 detector

Claims (11)

Detection means for detecting the object based on size information relating to the size of the object to be detected in the area from the area set for the image;
Updating means for updating the size information based on the size of the object detected by the detecting means;
An image processing apparatus. The detection means detects the object based on the size information of each of the plurality of regions from the plurality of regions set for the image,
The image processing apparatus according to claim 1, wherein the update unit updates the size information for each of the plurality of regions.   The image processing apparatus according to claim 1, wherein the detection unit detects the object based on the size information from the area in the image captured by the imaging unit.   The image processing apparatus according to claim 3, wherein when the angle of view of the imaging unit is changed, the update unit updates the size information based on the change of the angle of view of the imaging unit.   The updating unit is configured to select a current angle of view of the imaging unit among a plurality of pieces of size information corresponding to the plurality of field angles set by the imaging unit based on the size of the object detected by the detection unit. The image processing apparatus according to claim 3, wherein the size information corresponding to the number is updated. Further comprising a receiving means for receiving the designation of the area;
The image processing apparatus according to claim 1, wherein the detection unit detects the object based on the size information from an area where the designation is received by the reception unit. Further comprising setting means for setting the region for the image based on how the object appears in the image;
The image processing apparatus according to claim 1, wherein the detection unit detects the object from the area set for the image by the setting unit based on the size information.   The image processing apparatus according to claim 1, further comprising an output unit that outputs the size information in association with the area.   The image processing apparatus according to claim 8, wherein the output unit outputs each of the plurality of pieces of size information in association with each of the plurality of regions. An information processing method executed by an image processing apparatus,
A detection step of detecting the object based on the size information relating to the size of the object to be detected in the region from the region set for the image;
An update step of updating the size information based on the size of the object detected in the detection step;
An information processing method including:   A program for causing a computer to function as each unit of the image processing apparatus according to any one of claims 1 to 8.

Images