JP6098631B2 - Analysis processing device - Google Patents

Description

  The present invention relates to an analysis processing apparatus, and more particularly to an analysis processing apparatus that analyzes an image captured by a camera.

  In recent years, with the development of information processing technology, analysis systems for analyzing various data have been developed. For example, there is an analysis system that analyzes an image taken by a surveillance camera and detects the presence of a specific person or object registered in advance. In such an analysis system, it is required to detect the presence of a specific person or the like within a predetermined time after an image is taken by a surveillance camera. For this reason, it is desired to perform high-speed analysis processing.

  On the other hand, when there are many surveillance cameras and many people and objects are simultaneously shown in many images taken by these cameras, there are many processes for detecting a specific person or object, and the analysis processing device Processing requests for are concentrated. Then, there is a problem that high-speed analysis processing cannot be realized by the analysis processing apparatus, and it takes a long time to detect a specific person or object. Even if there is a single surveillance camera, if a large number of people and objects are shown in the image at the same time, it will take a long time to detect a specific person or object as described above. .

  Here, Patent Document 1 describes a technique for detecting a specific person from an image taken by a surveillance camera. Specifically, in the technique of Patent Literature 1, processing priority is set for each input image based on the position of the face image detected in the input image input from the camera. For example, it is described that, for an image in which a face region exists at a position close to the horizontal edge of the image, the priority of analysis processing for the image is set high.

JP 2011-70576 A

  However, in the technique disclosed in Patent Document 1 described above, the priority of analysis processing is set for each “input image” based on conditions such as the position of the face image detected in the input image. For this reason, an input image set with a high priority is preferentially processed. However, an analysis process is performed on the entire input image, and therefore an analysis process for an unnecessary image region can also be performed. . As a result, there arises a problem that high-speed analysis processing cannot be performed by the analysis processing apparatus.

  Therefore, an object of the present invention is to solve the problem that it takes time to perform analysis processing on an image in an analysis processing apparatus, a program, and an analysis processing method that perform image processing.

An analysis processing apparatus according to one aspect of the present invention is
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object position determination means for determining whether or not the object region is located in a predetermined specific region in the image;
Priority setting means for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination means;
With
The configuration is as follows.

Moreover, the program which is the other form of this invention is:
In the information processing device,
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object position determination means for determining whether or not the object region is located in a predetermined specific region in the image;
Priority setting means for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination means;
It is a program for realizing.

In addition, an analysis processing method according to another embodiment of the present invention includes:
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Set as
Determining whether the object region is located in a predetermined specific region in the image;
In accordance with a determination result of whether or not the object region is located in the specific region, a priority level representing a priority level for executing a predetermined analysis process on the object region is set.
The configuration is as follows.

  The present invention can provide an analysis processing device, a program, and an analysis processing method that can increase the speed of analysis processing for an image by being configured as described above.

It is a block diagram which shows the structure of the video analysis system in Embodiment 1 of this invention. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. It is a figure which shows an example of the data memorize | stored in the video analysis system disclosed in FIG. It is a figure which shows an example of the data memorize | stored in the video analysis system disclosed in FIG. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. It is a figure which shows an example of the data memorize | stored in the video analysis system disclosed in FIG. It is a figure which shows an example of the data memorize | stored in the video analysis system disclosed in FIG. It is a figure which shows operation | movement of the image | video analysis system disclosed in FIG. It is a flowchart which shows operation | movement of the image | video analysis system disclosed in FIG. It is a block diagram which shows the structure of the video analysis system in Embodiment 2 of this invention. It is a figure which shows an example of the data memorize | stored in the video analysis system disclosed in FIG. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. It is a figure which shows an example of the image processing by the video analysis system disclosed in FIG. FIG. 12 is a diagram illustrating an operation of the video analysis system disclosed in FIG. 11. 12 is a flowchart illustrating an operation of the video analysis system disclosed in FIG. 11. It is a block diagram which shows the structure of the video analysis system in additional remark 1-1 of this invention.

<Embodiment 1>
A first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a diagram illustrating a configuration of a video analysis system, and FIGS. 2 to 8 are diagrams illustrating data stored in the video analysis system and a state of analysis processing. 9 to 10 are diagrams illustrating the operation of the video analysis system.

  The video analysis system 10 according to the present embodiment is configured by one or a plurality of information processing apparatuses, and is an analysis processing apparatus that performs a predetermined analysis process on an image. Specifically, the video analysis system 10 performs processing for detecting a person registered in advance from an image captured by the monitoring camera 20 and notifying the alarm display application 30 of the detected person. Note that the video analysis system 10 is an example of an analysis processing apparatus according to the present invention, and is not necessarily limited to executing the analysis processing described below, and may perform other analysis processing.

  An example of the configuration of the video analysis system 10 is shown in FIG. As shown in this figure, the video analysis system 10 is first connected with a plurality of the monitoring cameras 20 described above. For example, each monitoring camera 20 is installed at a different location in a building to be monitored, and takes images (images) of each space such as each room in the building and transmits the images to the analysis system 10. To do.

  Further, the above-described alarm display application 30 is connected to the video analysis system 10. The alarm display application 30 is incorporated in, for example, an information processing terminal of a person who performs monitoring in a building. As will be described later, in order to notify the monitor of information notified from the video analysis system 10, notification information is provided. Is displayed on the screen of the information processing terminal, or sound or light is output.

  In addition, the video analysis system 10 is constructed by incorporating a program into the equipped arithmetic device, and includes a face position detection unit 11, a face feature analysis unit 12, a face collation unit 13, an analysis processing priority calculation unit 15, and an analysis. A processing priority control unit 16 is provided. In addition, the video analysis system 10 includes a face feature information database 14, an analysis priority shooting area information storage unit 17, and an analysis processing task group storage unit 18 in a storage device that is equipped. Hereinafter, each configuration will be described in detail.

  The face position detection unit 11 (object detection means) acquires the video imaged by each monitoring camera 20 and transmitted from each monitoring camera 20, and is displayed in the frame for each frame of the acquired video image. Detect human faces. Specifically, the face position detection unit 11 acquires frames constituting an image at a predetermined sampling period (arrow Y1 in FIG. 9 and step S1 in FIG. 10), and detects a human face from within each frame. For example, the position of the human face shown in the frame is detected based on preset information (color and shape information) representing the characteristics of the human face.

  Here, an example when a human face is detected in a frame for each surveillance camera 20 is shown in FIG. In FIG. 2, the upper diagram shows a state when a human face is detected in a frame acquired from the surveillance camera 20 with the camera ID “001” assigned in advance, and the lower diagram is assigned in advance at almost the same date and time. A state when a human face is detected in a frame acquired from the monitoring camera 20 with the camera ID “002” is shown. In the example of this figure, the human face indicated by reference signs P11, P12, and P13 is detected from the frame acquired from the monitoring camera 20 with the camera ID “001”, and the frame acquired from the monitoring camera 20 with the camera ID “002”. The human faces indicated by reference numerals P21 and P22 are detected.

  Then, the face position detection unit 11 sets an image area in a range including the face as a person area (object area) with reference to the detected position of the human face in the frame (step S2 in FIG. 10). . Specifically, the face position detection unit 11 sets a rectangular area surrounding the entire area detected as a human face as a person area, or a rectangular area set in advance around the center of gravity of the position detected as a human face. Is set as a person area.

  Here, FIG. 3 shows an example when a person area is set in the acquired frame. In FIG. 3, the upper diagram shows the situation when the person area is set in the monitoring camera 20 with the camera ID “001”, and the lower figure shows the situation when the person area is set in the monitoring camera 20 with the camera ID “002”. It shows a state. In the example of this figure, person areas A11, A12, and A13 corresponding to human faces indicated by reference signs P11, P12, and P13 are set from the frame acquired from the monitoring camera 20 with the camera ID “001”, and the camera ID is “ Person regions A21 and A22 corresponding to the human face indicated by symbols P21 and P22 are set from the frame acquired from the monitoring camera 20 of “002”.

  In the above description, an example of a method of detecting a human face and a method of setting a human area by the face position detection unit 11 has been described. However, these methods are not limited to the above-described methods, and other methods are used. May be. Moreover, although the case where a human face was detected was illustrated above, you may set the area | region (object area | region) of the said object by detecting another preset object instead of a human face.

  Then, the face position detection unit 11 transmits information on the person area set for each frame of the video captured by each monitoring camera 20 to the analysis processing priority calculation unit 15 (arrow Y2 in FIG. 9). At this time, the face position detection unit 11 includes a camera ID that identifies the captured surveillance camera 20, date information when the frame is captured, image information of the entire frame, and information that identifies the range of the person area. The data to be analyzed is transmitted to the analysis processing priority calculation unit 15.

  The analysis processing priority calculation unit 15 (object position determination unit, priority setting unit) receives the analysis target data transmitted from the face position detection unit 11 and calculates the priority of each person area. Specifically, the analysis processing priority calculation unit 15 first extracts the camera ID of the monitoring camera 20 taken from the analysis target data and information for specifying the range of the person area, and monitors the same as the extracted camera ID. The specific area information set for the camera 20 is acquired from the analysis priority imaging area information storage unit 17 (arrow Y3 in FIG. 9). Here, the specific area information is information indicating a predetermined specific area in a frame shot by the monitoring camera 20 and is set for each monitoring camera 20.

  For example, as shown in FIG. 4, for the monitoring area 20 with the camera ID “001” in the upper diagram, the specific area is set to a rectangular area indicated by reference numerals R11 and R12. For the monitoring camera 20 with "002", rectangular areas indicated by reference numerals R21 and R22 are set. As shown in FIG. 5, the information on the specific area is actually represented by the coordinates of the four corners of the specific area that is a rectangle, and the camera ID is associated with the preset priority, and analysis priority is given. It is stored in the shooting area information storage unit 17. The priority associated with each specific area is, for example, a value larger than 0, and a value preset by the administrator side according to the position of the specific area is stored. However, the priority associated with each specific area may be updated according to the result of the analysis process. Further, the specific area is not necessarily limited to a rectangle.

  Then, the analysis processing priority calculation unit 15 determines whether or not the person region transmitted from the face position detection unit 11 is located in a specific region set for the monitoring camera 20 that has captured the frame of the person region. (Step S3 in FIG. 10). At this time, the analysis processing priority calculation unit 15 determines that the person area is located in the specific area when at least a part of the person area is located in the specific area. Specifically, first, a camera ID in which a frame including a person area to be determined is captured is acquired, and coordinates of a specific area associated with the same camera ID as the camera ID are read out. Then, it is determined whether or not at least some of the coordinates of the person region that is the determination target are located within a specific region set for the same monitoring camera 20.

  Here, when the person area A11 and the like are set as shown in FIG. 3 and the specific area R11 and the like as shown in FIG. 4 are set, the relationship between the person area A11 and the specific area R11 is As shown in FIG. In this case, the person areas A11 and A13 in the frame acquired from the monitoring camera 20 with the camera ID “001” are positioned in the specific areas R11 and R12 associated with the monitoring camera 20 with the same camera ID “001”. It is determined that Further, it is determined that the person area A21 in the frame acquired from the monitoring camera 20 with the camera ID “002” is located in the specific area R21 associated with the monitoring camera 20 with the same camera ID “002”. . On the other hand, it is determined that the person areas A12 and A22 are not located in the specific area.

  And the analysis process priority calculation part 15 sets the priority showing the priority which performs an analysis process with respect to each person area | region based on the determination result mentioned above. Specifically, a priority based on a priority value associated with the specific area is set for a person area determined to be located in the specific area. For example, the priority of the person area is set by setting the priority value associated with the specific area as it is or the value obtained by multiplying the priority value by a predetermined weight as the priority of the person area. The degree is greater than zero. On the other hand, the priority is set to 0 for the person area determined not to be located in the specific area. Thereby, the priority value of the person area determined to be located in the specific area is set to be higher than the priority value of the person area determined not to be located in the specific area (step S4 in FIG. 10). ).

  After that, the analysis processing priority calculation unit 15 sets the priority set for the person area in the analysis target data such as the image information of the entire frame, the date / time information at the time of shooting, and the information specifying the range of the person area. Is sent to the analysis processing priority control unit 16 (arrow Y4 in FIG. 9).

  The analysis processing priority control unit 16 (analysis processing means) generates an analysis processing task based on the analysis target data received from the analysis processing priority calculation unit 15 and stores it in the analysis processing task group storage unit 18 ( Arrow Y5 in FIG. 9). Specifically, the analysis processing priority control unit 16 assigns a task ID to each person area of each analysis target data based on information on the analysis processing procedure set in advance and the received analysis target data, and performs analysis. The processing contents (analysis engine name), the priority, and the time information when the frame including the person area is captured are associated and stored as the analysis processing task group storage unit 18. Here, FIG. 7 shows an example of the analysis processing task stored in the analysis processing task group storage unit 18. In this figure, “analysis data” at the right end is image information located in the range of the person area, and the above-described imaged date / time information, priority, and analysis engine are associated with each other.

  In the present embodiment, the analysis process whose execution is controlled by the analysis process priority control unit 16 extracts the facial feature amount of the person shown in the person area executed by the face feature analysis unit 12. And a collation process for identifying a person by collating the extracted feature quantity of the person's face with a pre-registered feature quantity, which is executed by the face collating unit 13. The feature amount registered in advance is a feature amount of a person's face to be found by analysis, and the feature amount generated in advance is stored in the face feature information database 14. However, in the present invention, the analysis process whose execution is controlled by the analysis process priority control unit 16 is not limited to the analysis process described above.

  Then, the analysis processing priority control unit 16 sorts the table of analysis processing task groups generated and stored as shown in FIG. 7 in descending order of priority as shown in FIG. Then, the analysis processing priority control unit 16 performs control so as to execute the analysis processing for the person area according to the table of analysis processing task groups sorted in descending order of priority. That is, the analysis processing by the face feature analysis unit 12 and the face matching unit 13 is controlled so that the feature amount extraction processing and the matching processing are executed on the image information in the person area in descending order of the set priority (see FIG. 10 step S5, arrows Y5, Y6, Y7, Y8 in FIG. 9).

  Thereafter, the analysis processing priority control unit 16 notifies the alarm analysis application 30 of the final analysis result (arrow Y9 in FIG. 9). Here, when a person registered in the face feature information database is detected by the collation processing, information on the date, the camera ID of the monitoring camera 20, the coordinates in the frame, etc. are taken together with the person information. Notify information related to the recorded video.

  As a result, when a person's face is projected in the specific area, the analysis processing for the image information of the person area, which is the person's face area, is preferentially executed. Therefore, analysis processing in the vicinity of the specific area can be performed at high speed, and person detection can be performed quickly.

  In addition to the above configuration, the analysis processing priority calculation unit 15 may execute a process of updating the priority of the specific area stored in the analysis priority imaging area information storage unit 17. As an example, according to the determination result of whether or not the above-described person area is located in the specific area, the priority of the specific area depends on the number of times the person area is located in the specific area and the ratio to the whole area. An update may be performed to change the value of. Specifically, when the number of times is large or the ratio is large, the priority value may be changed to a higher value.

<Embodiment 2>
A second embodiment of the present invention will be described with reference to FIGS. FIG. 11 is a diagram illustrating a configuration of the video analysis system, and FIGS. 12 to 15 are diagrams illustrating data stored in the video analysis system and a state of analysis processing. 16 to 17 are diagrams illustrating the operation of the video analysis system.

  The video analysis system 10 in the present embodiment has almost the same configuration as that of the first embodiment described above. That is, the video analysis system 10 includes a face position detection unit 11, a face feature analysis unit 12, a face collation unit 13, an analysis processing priority calculation unit 15, and an analysis, which are constructed by incorporating a program into an equipped arithmetic device. A processing priority control unit 16 is provided. In addition, the video analysis system 10 includes a face feature information database 14, an analysis priority shooting area information storage unit 17, and an analysis processing task group storage unit 18 in a storage device that is equipped. In addition to the above configuration, the video analysis system according to the present embodiment includes a frame face position information storage unit 19 as shown in FIG. 11, and the function of the analysis processing priority calculation unit 15 is different. Hereinafter, the configuration different from the first embodiment will be mainly described in detail.

  The face position detection unit 11 (object detection means) in the present embodiment obtains an image captured by each monitoring camera 20 and transmitted from each monitoring camera 20 (arrow Y1 in FIG. 16, step S11 in FIG. 17). ) For each frame of the acquired video, the human face reflected in the frame is detected. The process for detecting a face is as described in the first embodiment.

  Then, the face position detection unit 11 sets, as a person area (object area), an image area in a range including the face with reference to the detected position of the human face in the frame (step S12 in FIG. 17). . The person area setting process is as described in the first embodiment.

  Then, the face position detection unit 11 transmits information on the person area set for each frame of the video captured by each monitoring camera 20 to the analysis processing priority calculation unit 15 (arrow 2 in FIG. 16). At this time, the face position detection unit 11 includes a camera ID that identifies the captured surveillance camera 20, date information when the frame is captured, image information of the entire frame, and information that identifies the range of the person area. Then, it is transmitted to the analysis processing priority calculation unit 15 as analysis target data. In addition, when a frame is acquired from each monitoring camera 20, the face position detection unit 11 sequentially identifies a person area as described above and transmits the person region to the analysis processing priority calculation unit 15.

  The analysis processing priority calculation unit 15 (object region information storage processing unit, new object region specifying unit) receives and stores the analysis target data transmitted from the face position detection unit 11 and also uses the same monitoring camera 20. A region (object region) that includes an object that has newly appeared in the past (for example, immediately before) frame captured in step (b) is specified as a new object region.

  Specifically, the analysis processing priority calculation unit 15 first receives the analysis target data transmitted from the face position detection unit 11, and stores the analysis target data in the frame face position information storage unit 19 (FIG. 16). Arrow Y3 ', step S13 in FIG. At this time, as shown in FIG. 12, the analysis target data stored includes, in the image information of the entire frame, information that identifies the camera ID of the captured surveillance camera 20, the range of the person area, and the captured date and time information. At least. The analysis processing priority calculation unit 15 stores the analysis target data in the frame face position information storage unit 19 every time the analysis target data is received from the face position detection unit 11.

  Subsequently, when the analysis processing priority calculation unit 15 newly receives the analysis target data from the face position detection unit 11, the analysis processing priority calculation unit 15 uses the same monitoring camera 20 as the monitoring camera 20 that captured the frame corresponding to the received analysis target data. Data to be analyzed relating to the captured past frame is read from the frame face position information storage unit 19 (arrow Y3 ′ in FIG. 16). For example, based on the camera ID and shooting date / time included in the new analysis target data, the analysis target data of the immediately previous frame shot by the same monitoring camera 20 is read.

  Here, for example, as shown in FIG. 13B, the analysis processing priority calculation unit 15 newly receives analysis target data of a frame shot by the monitoring camera 20 with the camera ID “001”. Assume that the analysis target data of the frame captured immediately before by the monitoring camera 20 with the same camera ID “001” is read from the frame face position information storage unit 19 as shown in FIG. In the new frame shown in FIG. 13B, person areas A41, A42, A43, and A44 corresponding to the human face are set, and in the frame immediately before (past) shown in FIG. Person areas A31, A32, A33 corresponding to human faces are set.

  Subsequently, the analysis processing priority calculation unit 15 determines that the number of person areas in the new frame is larger than the number of person areas in the immediately preceding frame (Yes in step S14 in FIG. 17). It is determined that a new face area has been added. Then, the analysis processing priority calculation unit 15 associates the person areas A41 to A44 in the new frame with the person areas A31 to A33 in the immediately preceding frame. At this time, for example, the person areas A41 to A44 in the new frame with a short distance between the person areas are associated with the person areas A31 to A33 in the immediately preceding frame.

  Specifically, the association processing by the analysis processing priority calculation unit 15 will be described with reference to FIG. FIG. 14A shows a person area (solid line) (FIG. 13B) in a new frame photographed by the same monitoring camera 20, and a person area (dotted line) (FIG. 13A) in the immediately preceding frame. , And FIG. 14B shows a state of processing for associating each person area. As shown in FIG. 14B, from the upper left corner of the rectangle which is one point of the person area (solid line) of the new frame to the upper left corner of the rectangle which is one point of the person area (dotted line) of the previous frame The distance is checked (step S15 in FIG. 17), and the person regions having the shortest distance are associated with each other (step S16 in FIG. 17). Thereafter, the distance between the remaining person areas is checked, the person areas with the shortest distance are attached, and this process is repeated (steps S15 and S16 in FIG. 17 are repeated). Here, the person area A31 in the immediately preceding frame is associated with the person area A41 in the new frame, and the person areas A32 and A42 and the person areas A33 and A43 are associated with each other. Then, the person area A44 in the new frame is not associated with any person area in the immediately preceding frame.

  Furthermore, the analysis processing priority calculation unit 15 examines the distances from the person area A44 in the new frame that is not associated to all the person areas A31, A32, and A33 in the immediately preceding frame. Then, the analysis processing priority calculation unit 15 is associated when the distance from the person area A44 in the new frame that has not been associated to the person area A31 in the immediately preceding frame is equal to or greater than a preset value. The missing person area A44 is specified as a new object area (step S17 in FIG. 17). Here, the new object region is an object region including a person who newly appears in a new frame with respect to the previous frame.

  The analysis processing priority calculation unit 15 determines the new object region A44 as the new object region A44 without examining the distances from the person region A44 to all the person regions A31, A32, A33 in the immediately preceding frame. May be specified. In addition, as described above, the analysis processing priority calculation unit 15 does not associate each person area with the shortest distance, and associates each person area with a distance shorter than a preset distance. In addition, the person area that is not associated may be specified as the new object area.

  Here, the method by which the analysis processing priority calculation unit 15 specifies the new object region is not limited to the method described above. For example, the distance between the person areas calculated when associating the person areas may be the distance between the centroids of the person areas, or the person areas may be associated with each other based on other criteria. Good. For example, comparing the area of the person area in the new frame and the area of the person area in the previous frame without using the distance between the person areas, Corresponding may be performed, and those having a difference of each area smaller than a certain value may be associated.

  Also, in the above description, as a method for specifying a new object region in a new frame, the use of the immediately preceding frame with respect to the new frame has been described. The frame may be used. For example, the new object area in the new frame may be specified by associating the person area in the new frame with the person area in one or a plurality of frames in the past preset time.

  Subsequently, the analysis processing priority calculation unit 15 (object position determination means, priority setting means) calculates the priority of the person area specified as the new object area. Specifically, the analysis processing priority calculation unit 15 specifies the specific area set for the same monitoring camera 20 from the camera ID of the monitoring camera 20 in which the frame including the person area specified as the new object area is captured. Information is acquired from the analysis priority imaging region information storage unit 17 (arrow Y3 in FIG. 16). Here, the specific area information is information indicating a predetermined specific area in a frame shot by the monitoring camera 20 and is set for each monitoring camera 20. The specific area is the same as that described in the first embodiment, and detailed description thereof is omitted.

  Then, the analysis processing priority calculation unit 15 determines whether or not the person area specified as the new object area is located in the specified area set for the monitoring camera 20 that captured the frame of the person area. At this time, the analysis processing priority calculation unit 15 determines that the person area is located in the specific area when at least a part of the person area is located in the specific area.

  Here, the upper diagram in FIG. 15 shows the positions of the specific areas R11 and R12 (dotted lines) set for the monitoring camera 20 with the camera ID “001”, and the frames acquired from the same monitoring camera 20 are shown. In the figure, a person area A44 is identified as a new object area. In this case, it is not determined that the person area A44 specified as the new object area is located in the specific area. The lower diagram of FIG. 15 shows the positions of the specific areas R21 and R22 (dotted lines) set for the monitoring camera 20 with the camera ID “002”. In addition, in the frame acquired from the same monitoring camera 20 The person area A51 is identified as a new object area. In this case, it is determined that the person area A51 specified as the new object area is located in the specific area.

  And the analysis process priority calculation part 15 sets the priority showing the priority which performs an analysis process with respect to each person area | region based on the determination result mentioned above. Specifically, a higher priority value than the other person areas is set for the person area specified as the new object area. For example, the priority of the person area specified as the new object area is set to a value greater than 0, and the priorities of the other person areas are set to 0. Further, among the person areas specified as the new object area, a higher priority value is set for the person area determined to be located in the specific area as described above. As a result, a higher priority is set in the order of the person area that is the new object area and located in the specific area, the person area that is the new object area but is not located in the specific area, and the other person areas (FIG. 17). Step S17).

  After that, the analysis processing priority calculation unit 15 sets the priority set for the person area in the analysis target data such as the image information of the entire frame, the date / time information at the time of shooting, and the information specifying the range of the person area. Is sent to the analysis processing priority control unit 16 (arrow Y4 in FIG. 16).

  Similar to the first embodiment, the analysis processing priority control unit 16 (analysis processing unit) generates an analysis processing task based on the analysis target data received from the analysis processing priority calculation unit 15, and generates an analysis processing task group. It memorize | stores in the memory | storage part 18 (arrow Y5 of FIG. 16). Then, the analysis processing priority control unit 16 sorts the table of the analysis processing task group in descending order of priority, and performs control so as to execute the analysis processing on the person region in descending order of priority. That is, the analysis processing by the face feature analysis unit 12 and the face matching unit 13 is controlled so that the feature amount extraction processing and the matching processing are executed on the image information in the person area in descending order of the set priority (see FIG. 17 (step S18, arrows Y5, Y6, Y7, Y8 in FIG. 16).

  Thereafter, the analysis processing priority control unit 16 notifies the final analysis result to the alarm display application 30 (arrow Y9 in FIG. 16). Here, when a person registered in the face feature information database is detected by the collation processing, information on the date, the camera ID of the monitoring camera 20, the coordinates in the frame, etc. are taken together with the person information. Notify information related to the recorded video.

  As a result, when a new human face appears in a new frame, the analysis processing for the image information of the human area, which is the human face area, is preferentially executed. In particular, when a person area newly appearing in a new frame is located in a specific area, the analysis process for the image information of the person area is further preferentially executed. For this reason, analysis processing of a required image area can be performed at high speed, and person detection can be performed quickly.

  Here, the above-described program is stored in a storage device or recorded on a computer-readable recording medium. For example, the recording medium is a portable medium such as a flexible disk, an optical disk, a magneto-optical disk, and a semiconductor memory.

  Although the present invention has been described with reference to the above-described embodiment and the like, the present invention is not limited to the above-described embodiment. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

<Appendix>
Part or all of the above-described embodiment can be described as in the following supplementary notes. The outline of the configuration of the analysis processing apparatus (see FIG. 18), program, and analysis processing method in the present invention will be described below. However, the present invention is not limited to the following configuration.

(Appendix 1-1)
An object detection unit 101 that acquires an image captured by the camera 110 and detects an object having a preset feature in the acquired image;
A predetermined range based on the position where the object is detected by the object detection means 101 in the image is set as an object region, and whether or not the object region is located in a predetermined specific region in the image Object position determination means 102 for determining whether or not
A priority setting unit 103 for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination unit 102;
The analysis processing apparatus 100 provided with.

(Appendix 1-2)
The analysis processing device according to attachment 1-1,
The priority setting means is configured such that the priority of the object area located in the specific area is higher than the priority of the object area not located in the specific area according to the determination result by the object position determination means. Set,
Analysis processing device.

(Appendix 1-3)
The analysis processing device according to appendix 1-1 or 1-2,
The priority setting means sets the priority of the object region located in the specific region based on a priority value preset for each specific region;
Analysis processing device.

(Appendix 1-4)
An analysis processing apparatus according to any one of appendices 1-1 to 1-3,
The object detection means acquires images captured by a plurality of cameras, detects an object having a preset feature in the images acquired for each camera,
The object position determination means sets the object region in the image captured for each camera, and the object region is the same camera as the camera from which the image in which the object region is set is captured. It is determined whether or not it is located in the specific area set in advance,
The priority setting unit is configured such that the priority of the object region located in the specific region with respect to the same camera is not located in the specific region with respect to the same camera according to a determination result by the object position determination unit. Set to be higher than the priority of the object area,
Analysis processing device.

(Appendix 1-5)
The analysis processing apparatus according to appendix 1-4,
The priority setting means sets the priority of the object area located in the specific area set for the camera based on a priority value set in advance for each camera and for each specific area.
Analysis processing device.

(Appendix 1-6)
An analysis processing apparatus according to any one of appendices 1-1 to 1-5,
An analysis processing means for performing a preset analysis process on the data in the object region in descending order of the priority values set in the object region by the priority setting unit;
Analysis processing device.

(Appendix 1-7)
The analysis processing device according to attachment 1-6,
The analysis processing means includes, as the preset analysis process, a feature quantity extraction process for extracting a preset feature quantity from the data in the object region, and the extracted feature quantity and a preset collation target And a matching process for matching with the feature quantity
Analysis processing device.

(Appendix 1-8)
In the information processing device,
An object detection means for acquiring an image captured by the camera and detecting an object having a preset feature in the acquired image;
A predetermined range based on the position where the object is detected by the object detection means in the image is set as an object area, and whether or not the object area is located in a predetermined specific area in the image Object position determination means for determining
Priority setting means for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination means;
A program to realize

(Appendix 1-9)
The program according to appendix 1-8,
The priority setting means is configured such that the priority of the object area located in the specific area is higher than the priority of the object area not located in the specific area according to the determination result by the object position determination means. Set,
program.

(Appendix 1-10)
Acquire an image captured by the camera, and detect an object having a preset feature in the acquired image,
A predetermined range based on the position where the object is detected in the image is set as an object region, and it is determined whether the object region is located in a predetermined specific region in the image,
In accordance with a determination result of whether or not the object region is located in the specific region, a priority level representing a priority level for executing a predetermined analysis process on the object region is set.
Analysis processing method.

(Appendix 1-11)
The analysis processing method according to attachment 1-10, wherein:
When setting the priority, the priority of the object region located in the specific region is not located in the specific region according to a determination result of whether or not the object region is located in the specific region. Set to be higher than the priority of the object area,
Analysis processing method.

(Appendix 2-1)
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object area information storage processing means for storing information of the object area in a storage device;
Based on the information of the object area stored in the storage device in the image acquired in the past and the information of the object area set by the object detection means in the newly acquired image, New object region specifying means for specifying, as a new object region, the object region that includes a newly appearing object in the image newly acquired in the image acquired in the past;
An analysis processing apparatus comprising:

(Appendix 2-2)
An analysis processing apparatus according to appendix 2-1,
The new object area specifying means includes information on the object area set by the object detection means in the newly acquired image, and the object area stored in the storage device in the image acquired immediately before And identifying the object region including an object that newly appears in the newly acquired image as the new object region based on the information of
Analysis processing device.

(Appendix 2-3)
The analysis processing device according to attachment 2-1 or 2-2,
The new object area specifying means includes information on the object area set by the object detection means in the newly acquired image, and the object area stored in the storage device in the image acquired in the past. And the object region in the newly acquired image is associated with the object region in the image acquired in the past according to a preset criterion based on the information of Identifying the object region in the newly acquired image that is not associated with the object region in the image as the new object region;
Analysis processing device.

(Appendix 2-4)
The analysis processing device according to attachment 2-3,
The new object area specifying unit performs association between the object areas based on a distance of the object area in the newly acquired image with respect to the object area in the image acquired in the past.
Analysis processing device.

(Appendix 2-5)
An analysis processing apparatus according to appendix 2-3 or 2-4,
The new object region specifying means is newly acquired in which a distance of the object region in the newly acquired image is greater than or equal to a preset value with respect to the object region in the image acquired in the past. Identifying the object region in the image as the new object region;
Analysis processing device.

(Appendix 2-6)
The analysis processing device according to any one of supplementary notes 2-1 to 2-5,
Analysis processing means for performing pre-set analysis processing on data in the object area specified as the new object area by the new object area specifying means in preference to data in other object areas With
Analysis processing device.

(Appendix 2-7)
The analysis processing device according to any one of supplementary notes 2-1 to 2-5,
The new object area specifying means has a priority indicating the priority of executing a predetermined analysis process on the object area specified as the new object area by the new object area specifying means. Set to be higher than
An analysis processing means for performing a preset analysis process on the data in the object area in descending order of the priority set in the object area by the new object area specifying means,
Analysis processing device.

(Appendix 2-8)
The analysis processing device according to appendix 2-6 or 2-7,
The analysis processing means includes, as the preset analysis process, a feature quantity extraction process for extracting a preset feature quantity in the object region, and the extracted feature quantity and a feature to be a preset collation target A collation process for collating with the quantity,
Analysis processing device.

(Appendix 2-9)
In the information processing device,
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object area information storage processing means for storing information of the object area in a storage device;
Based on the information of the object area stored in the storage device in the image acquired in the past and the information of the object area set by the object detection means in the newly acquired image, New object region specifying means for specifying, as a new object region, the object region that includes a newly appearing object in the image newly acquired in the image acquired in the past;
A program to realize

(Appendix 2-10)
A program according to appendix 2-9,
The new object area specifying means includes information on the object area set by the object detection means in the newly acquired image, and the object area stored in the storage device in the image acquired immediately before And identifying the object region including an object that newly appears in the newly acquired image as the new object region based on the information of
program.

(Appendix 2-11)
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Set as
Storing information of the object region in a storage device;
Based on the information on the object region stored in the storage device in the image acquired in the past and the information on the object region set in the newly acquired image, the information acquired in the past Specifying the object region including a newly appearing object in the image newly acquired in the image as a new object region;
Analysis processing method.

(Appendix 2-12)
The analysis processing method according to attachment 2-11,
When specifying the new object region, information on the object region set in the newly acquired image, information on the object region stored in the storage device in the image acquired immediately before, Based on the above, the object region that includes a newly appearing object in the image that is newly acquired relative to the image acquired immediately before is specified as the new object region.
Analysis processing method.

  The present invention enjoys the benefit of the priority claim based on the patent application of Japanese Patent Application No. 2012-030466, filed on February 15, 2012 in Japan, and is described in the patent application. The contents are all included in this specification.

DESCRIPTION OF SYMBOLS 10 Image | video analysis system 11 Face position detection part 12 Face feature analysis part 13 Face collation part 14 Face feature information database 15 Analysis process priority calculation part 16 Analysis process priority control part 17 Analysis priority imaging | photography area information storage part 18 Analysis process task group Storage unit 19 Frame face position information storage unit 20 Surveillance camera 30 Alarm display application 100 Analysis processing device 101 Object detection unit 102 Object position determination unit 103 Priority setting unit 110 Camera

Claims (10)

An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object position determination means for determining whether or not the object region is located in a predetermined specific region in the image;
Priority setting means for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination means ;
Bei to give a,
The object detection means acquires images captured by a plurality of cameras, detects an object having a preset characteristic in the acquired image for each camera, and detects the detected object. Set the object area
The object position determination means determines whether the object region is located in the specific region set in advance for the same camera as the camera from which the image in which the object region is set is captured,
The priority setting unit is configured such that the priority of the object region located in the specific region with respect to the same camera is not located in the specific region with respect to the same camera according to a determination result by the object position determination unit. Set to be higher than the priority of the object area,
Analysis processing device. The analysis processing device according to claim 1,
The priority setting means is configured such that the priority of the object area located in the specific area is higher than the priority of the object area not located in the specific area according to the determination result by the object position determination means. Set,
Analysis processing device. The analysis processing device according to claim 1 or 2,
The priority setting means sets the priority of the object region located in the specific region based on a priority value preset for each specific region;
Analysis processing device. The analysis processing device according to any one of claims 1 to 3 ,
The priority setting means sets the priority of the object area located in the specific area set for the camera based on a priority value set in advance for each camera and for each specific area.
Analysis processing device. The analysis processing device according to any one of claims 1 to 4,
An analysis processing means for performing a preset analysis process on the data in the object region in descending order of the priority values set in the object region by the priority setting unit;
Analysis processing device. The analysis processing device according to claim 5 ,
The analysis processing means includes, as the preset analysis process, a feature quantity extraction process for extracting a preset feature quantity from the data in the object region, and the extracted feature quantity and a preset collation target And a matching process for matching with the feature quantity
Analysis processing device. In the information processing device,
An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Object detection means to be set as
Object position determination means for determining whether or not the object region is located in a predetermined specific region in the image;
Priority setting means for setting a priority indicating a priority level for executing a predetermined analysis process on the object region in accordance with a determination result by the object position determination means ;
And realize
The object detection means acquires images captured by a plurality of cameras, detects an object having a preset characteristic in the acquired image for each camera, and detects the detected object. Set the object area
The object position determination means determines whether the object region is located in the specific region set in advance for the same camera as the camera from which the image in which the object region is set is captured,
The priority setting unit is configured such that the priority of the object region located in the specific region with respect to the same camera is not located in the specific region with respect to the same camera according to a determination result by the object position determination unit. Set to be higher than the priority of the object area,
A program to make things happen. The program according to claim 7,
The priority setting means is configured such that the priority of the object area located in the specific area is higher than the priority of the object area not located in the specific area according to the determination result by the object position determination means. Set,
program. An image captured by the camera is acquired, an object having a preset characteristic is detected in the acquired image, and a predetermined range based on a position in the image where the object is detected is defined as an object region. Set as
Determining whether the object region is located in a predetermined specific region in the image;
In accordance with a determination result of whether or not the object region is located in the specific region, setting a priority that represents a priority for performing a predetermined analysis process on the object region ,
When setting the object area, the image acquired by each of the plurality of cameras is acquired, and an object having a preset characteristic in the acquired image is detected for each camera, and the detection is performed. Set the object area of the object,
When determining whether or not the object area is located in the specific area, the object area is specified in advance for the same camera as the camera from which the image in which the object area is set is captured. Determine whether it is located in the area,
When setting the priority, the priority of the object region located in the specific region is the same for the same camera according to the determination result of whether or not the object region is located in the specific region. Setting the camera to be higher than the priority of the object area not located in the specific area
Analysis processing method. The analysis processing method according to claim 9, wherein
When setting the priority, the priority of the object region located in the specific region is not located in the specific region according to a determination result of whether or not the object region is located in the specific region. Set to be higher than the priority of the object area,
Analysis processing method.

Images