JP6627365B2 - Information processing method, information processing apparatus, and program - Google Patents

Description

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

  It is possible to derive the trajectory of a person using images captured by a plurality of surveillance cameras whose imaging regions do not overlap each other. The derived movement trajectory can be used for, for example, marketing.

  FIG. 1 is a diagram for explaining a method of deriving a movement trajectory using a plurality of monitoring cameras. On the upper side of FIG. 1, three surveillance cameras A, B and C are shown. Further, the imaging area of each camera is indicated by a broken line. Further, an image including the person Y, captured by the camera A, an image including the person X, captured by the camera B, and an image including the person Z, captured by the camera C, are illustrated.

  In the lower part of FIG. 1, the appearance times of the persons X, Y, and Z in the imaging region of each camera and a plurality of types of feature amounts extracted from the image of each camera are shown in a table format. . Note that the first row of the table is information obtained from the image captured by the camera B, the second row is information obtained from the image captured by the camera A, and the third row is the camera C is information obtained from the image captured by C.

  By comparing the feature values of each row in the table, it is possible to determine whether each person is different. FIG. 1 shows an example in which the person X and the person Y are determined to be the same person. As a result, it is derived that the movement trajectory of the person is as shown by the upper arrow in FIG.

JP 2010-231254 A WO 2014/050432 JP-A-2004-72628

  In the case of constructing the system as described above, if a certain computer performs processing based on an image captured by each monitoring camera, the image captured by each monitoring camera is transmitted to the computer. There is a need. Therefore, the load on the network increases.

  Therefore, it is conceivable to arrange a processing device for each monitoring camera and distribute the extraction of the feature amount and the determination of the difference between the persons to the respective processing devices.

  FIG. 2 is a diagram for explaining extraction of a feature amount and determination of a person's dissimilarity by a processing device for each monitoring camera. In FIG. 2, it is assumed that the person H has moved from the camera A to the camera B.

  (1) When the person H appears in the imaging region of the camera A, the processing device A detects the person H from an image captured by the camera A, and extracts a feature amount of the person H. The processing device A registers the feature amount as a feature amount of the person X in a DB (Data Base).

  (2) Thereafter, when the person H appears in the imaging region of the camera B, the processing device B detects the person H from the image captured by the camera B, and extracts the feature amount of the person H. The processing device B registers the feature amount in the DB as the feature amount of the person Y.

  (3) Subsequently, the processing device B transmits the feature amount of the person Y to the processing device A.

  (4) The processing device A compares the feature amount of the person Y transmitted from the camera B with the feature amount of the person X in the DB, and determines that the person Y and the person X are the same person. I do.

  (5) The processing device A transmits, to the processing device B, information indicating that the person Y and the person X are the same person as the determination result.

  According to the above configuration, it is not necessary to transmit an image captured by each surveillance camera via a network, so that a network load can be reduced.

  However, in an environment where a large number of people exist and a large number of surveillance cameras are installed, all the feature amounts (face, upper body color, lower body color, etc.) that can be extracted for each person are exchanged between the processing devices. In such a case, the total amount of the transmitted feature amounts increases.

  Therefore, in one aspect, an object of the present invention is to reduce the amount of data transmitted to derive a trajectory of a subject based on images captured by imaging devices installed at a plurality of locations.

In one proposal, an information processing device arranged for each of the imaging devices installed at a plurality of locations includes a first subject detected from an image captured by a first imaging device corresponding to the information processing device. Extracting a plurality of types of feature amounts, and for each of the plurality of types of feature amounts extracted in the extracting process, determining suitability for determining whether a subject captured by the first imaging device is different. A transmitting unit that selects a part of the characteristic amount having relatively high suitability and transmits information identifying the part of the characteristic amount selected in the selecting process to another information processing apparatus; Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the information processing device, the feature amount indicated by the information transmitted by the transmission unit is received. The judges a portion of selected features in the process of selecting, on the basis of the characteristic quantity received in the process of the receiving, the difference between the second object and the first object, the process Execute

  According to one aspect, it is possible to reduce the amount of data transmitted to derive a movement trajectory of a subject based on images captured by imaging devices installed at a plurality of locations.

It is a figure for explaining the derivation method of the movement locus using a plurality of surveillance cameras. FIG. 9 is a diagram for describing extraction of a feature amount and determination of a person's dissimilarity by a processing device for each monitoring camera. FIG. 1 is a diagram illustrating a configuration example of a monitoring system according to an embodiment of the present invention. FIG. 4 is a diagram for describing an outline of a process executed by the information processing device according to the embodiment of the present invention. FIG. 2 is a diagram illustrating a hardware configuration example of an information processing apparatus according to the embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a functional configuration of the information processing apparatus according to the embodiment of the present invention. It is a flowchart for demonstrating an example of the processing procedure of a feature-value extraction process. FIG. 3 is a diagram illustrating a configuration example of a person information storage unit. It is a flowchart for demonstrating an example of the processing procedure of transmission processing of selection characteristic information. It is a figure showing an example of a change of a situation according to a time zone about an imaging field of the same camera. It is a flowchart for demonstrating an example of the processing procedure of the reception processing of selection characteristic information. FIG. 5 is a diagram illustrating a configuration example of a selected feature storage unit. It is a flowchart for demonstrating an example of the processing procedure of the transmission process of the same person determination request. It is a flowchart for demonstrating an example of the processing procedure of the same person determination process.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 3 is a diagram illustrating a configuration example of the monitoring system according to the embodiment of the present invention. 3, a monitoring system 1 includes a plurality of cameras 20, such as a camera 20a, a camera 20b, a camera 20c, and a camera 20d, and information processing apparatuses 10a, 10b, 10c, and 10d. And a plurality of information processing apparatuses 10. Note that the camera 20a, the camera 20b, the camera 20c, and the camera 20d are referred to as the camera 20 when not distinguished. Similarly, when the information processing device 10b, the information processing device 10c, and the information processing device 10d are not distinguished, they are referred to as the information processing device 10.

  The cameras 20 are, for example, surveillance cameras 20 installed at a plurality of places in a city. The imaging areas of the cameras 20 are set so as not to overlap each other. However, a part of the imaging area of each camera 20 may overlap. Further, five or more cameras 20 may be installed, or three or less cameras 20 may be installed.

  Each camera 20 is connected to the corresponding information processing device 10. The information processing device 10 may be installed in a town together with the camera 20, for example. Each of the information processing apparatuses 10 exchanges a feature amount of a person included in an image captured by the camera 20 connected to the information processing apparatus 10 with another information processing apparatus 10, whereby the camera 20 The same person as the person imaged by another camera 20 is determined from the group of people imaged by the camera. By doing so, the movement trajectory of each person can be derived.

  FIG. 4 is a diagram for describing an outline of processing executed by the information processing device according to the embodiment of the present invention. FIG. 4 shows a method of finding the same person as the person imaged by the camera 20a from the group of people imaged by the camera 20b.

  First, for each person included in the image captured by the camera 20b, the information processing device 10b extracts a plurality of types of feature amounts from the image (S1). In the example of FIG. 4, an example is shown in which feature amounts relating to the color of the upper body, the color of the lower body, shoes, hair, and the like are extracted.

  Subsequently, the information processing device 10b maps a plurality of types of feature amounts to respective feature spaces to calculate a variance (S2). The information processing device 10b selects a part of the feature amount having a relatively large variance, and transmits information identifying the part of the feature amount to the information processing device 10a (S3). In FIG. 4, the color of the upper body is selected. Hereinafter, the selected feature is referred to as “selected feature”, and information for identifying the selected feature is referred to as “selected feature information”.

  The information processing device 10a transmits a feature value related to the color of the upper body to the information processing device 10b from among a plurality of types of feature values extracted for a person detected from an image captured by the camera 20a (S5). ).

  The information processing device 10b compares the feature value (upper body color) transmitted from the information processing device 10a with the feature value (upper body color) of each person extracted in the information processing device 10b, and the camera 20b The same person as the person imaged by the camera 20a is found from the group of imaged people.

  If the same person is not found, the information processing device 10a executes step S5 for another information processing device 10. In this case, the feature value transmitted in step S5 is selected based on the selected feature information notified in step S3 executed by the other information processing device 10.

  As described above, in the present embodiment, in the information processing apparatus 10 (in FIG. 4, the information processing apparatus 10b) that performs the determination of the person's dissimilarity, the feature amount determined to have a relatively high variance is determined by another The information is transmitted from the information processing device 10 to the information processing device 10. Therefore, the amount of data to be transmitted can be reduced as compared with the case where all of a plurality of types of feature amounts are transmitted. In addition, it is considered that a feature amount having a relatively high variance has a relatively high suitability for discriminating a person (a degree of discrimination of a person or a degree of contribution to discrimination of a person). In other words, it is considered that the feature amount that is not transmitted is a feature amount having low suitability for discriminating a person. Therefore, it is possible to suppress deterioration of the accuracy of determining the difference between persons due to partial transmission of the feature amount.

  Note that the selected feature amount is selected for each camera 20. This is because, since the imaging region of each camera 20 is different, a feature amount having relatively high suitability for discriminating a person is considered to be different for each camera 20. For example, it is difficult to determine the color of the upper body from the image captured by the camera 20a, or it is difficult to determine the shoes from the image captured by the camera 20b. This is because it is considered that the characteristics of the image to be obtained may be different.

  Hereinafter, the information processing apparatus 10 will be specifically described. FIG. 5 is a diagram illustrating a hardware configuration example of the information processing apparatus according to the embodiment of the present invention. The information processing apparatus 10 in FIG. 5 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like, which are mutually connected by a bus B.

  A program for realizing the processing in the information processing device 10 is provided by the recording medium 101. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not always be installed from the recording medium 101, and may be downloaded from another computer via a network. The auxiliary storage device 102 stores installed programs and also stores necessary files and data.

  The memory device 103 reads out the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes functions related to the information processing device 10 according to a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  Note that examples of the recording medium 101 include a portable recording medium such as a CD-ROM, a DVD disk, or a USB memory. An example of the auxiliary storage device 102 is a hard disk drive (HDD) or a flash memory. Each of the recording medium 101 and the auxiliary storage device 102 corresponds to a computer-readable recording medium.

  FIG. 6 is a diagram illustrating a functional configuration example of the information processing apparatus according to the embodiment of the present invention. 6, the information processing apparatus 10 includes a person detection unit 111, a feature amount extraction unit 112, a person tracking unit 113, a selected feature amount determination unit 114, a selected feature transmission unit 115, a selected feature reception unit 116, and a feature amount acquisition unit 117. , A determination request transmission unit 118, a determination request reception unit 119, an identical person determination unit 120, a determination result transmission unit 121, a determination result reception unit 122, and the like. Each of these units is realized by a process of causing the CPU 104 to execute one or more programs installed in the information processing apparatus 10. The information processing device 10 also uses a personal information storage unit 131 and a selected feature storage unit 132. These storage units can be realized using, for example, a storage device that can be connected to the auxiliary storage device 102 or the information processing device 10 via a network.

  The person detection unit 111 detects a person from subjects of a video (moving image) captured by the camera 20. When a plurality of persons are included in the video, each of the plurality of persons is detected. The feature amount extraction unit 112 extracts a plurality of types of feature amounts from the video of each person detected by the person detection unit 111. The person tracking unit 113 tracks each person in the video imaged by the camera 20. The tracking of the person is performed until the person leaves the imaging area of the camera 20. That is, tracking refers to tracking the movement of the same person in the video imaged by one camera 20. The person tracking unit 113 stores a plurality of types of feature amounts extracted for each tracked person in the person information storage unit 131.

  The selected feature amount determination unit 114 determines a selected feature amount from a plurality of types of feature amounts of a plurality of persons stored in the person information storage unit 131. The selected feature transmitting unit 115 transmits selected feature information indicating the selected feature amount to another information processing device 10.

  The selected feature receiving unit 116 receives selected feature information transmitted from another information processing device 10. The selected feature receiving unit 116 stores the received selected feature information in the selected feature storage unit 132 for each of the other information processing devices 10.

  The feature amount acquiring unit 117 acquires, from the person information storage unit 131, the feature amount of a person to be determined as the same person. The determination request transmission unit 118 transmits a request to determine whether or not there is the same person as the person related to the feature amount acquired by the feature amount acquisition unit 117 to another information processing apparatus 10. The determination request includes the feature amount indicated by the selected feature information for the other information processing device 10.

  The determination request receiving unit 119 receives a determination request for the same person transmitted from another information processing apparatus 10. The same person determination unit 120 compares and matches the feature amount included in the determination request received by the determination request receiving unit 119 with the feature amount stored in the personal information storage unit 131, and determines The presence or absence of the same person as the person is determined. The determination result transmission unit 121 transmits the determination result by the same person determination unit 120 to another information processing apparatus 10 that has transmitted the feature amount.

  The determination result receiving unit 122 receives a determination result for the same person determination request transmitted from another information processing apparatus 10.

  Hereinafter, a processing procedure executed by each information processing apparatus 10 will be described. FIG. 7 is a flowchart illustrating an example of a processing procedure of a feature amount extraction process.

  In step S101, the person detection unit 111 converts a video (moving image) input from the camera 20 into an image (still image), and detects a person from subjects in the image. For example, a video is converted into an image for each frame, and a person is detected from the image. With the detection of the person, the position information of the person in the image is specified. Further, time information indicating the current time is associated with the position information. The position information of the person may be a coordinate value in the image of any one of the pixels related to the person (for example, a pixel at the top (top)). Steps S101 to S104 are executed for each frame. A frame to be processed is called a “target frame”.

  Subsequently, the feature amount extraction unit 112 extracts a plurality of types of feature amounts from the target frame for each person detected in the target frame (S102). For example, feature quantities relating to the color of the upper body of the person, the color of the clothes of the lower body, the body type (height, width, etc.) of the person, etc. are extracted. For example, if the color is the upper body or the lower body, a value obtained by averaging the RGB components of the image corresponding to the upper body part or the lower body part may be extracted as the feature amount.

  Subsequently, the person tracking unit 113 detects the position information, time information, and each feature amount of each person detected from the target frame and the frame immediately before the target frame (hereinafter, referred to as “previous frame”). The position information, the time information, and the feature amount of each person are associated with each other by the same person and stored in, for example, the memory device 103, thereby tracking each person in the previous frame (S103). The same person is determined based on a plurality of types of feature amounts. For the detection and tracking of a person (S101 to S103), for example, a technique described in JP-A-2002-157599 may be used. Or "Takayoshi Yamashita and Hironori Fujiyoshi," Cooperative object tracking using the relationship between tracking target and surrounding area ", Image Recognition and Understanding Symposium (MIRU2011), pp. 56-63, 2011." The technique used may be used.

  Subsequently, the person tracking unit 113 determines whether or not the person being tracked has moved out of the imaging area of the camera 20 (S104). Specifically, when a person detected in the previous frame is not detected in the current frame, it is determined that the person has moved out of the imaging region. If the corresponding person does not exist (No in S104), step S101 and subsequent steps are repeated for each frame.

  If the corresponding person exists (Yes in S104), the person tracking unit 113 stores the history of the position information, the history of the time information, and the feature amounts stored in the memory device 103 for each person in each frame. The information is stored in the information storage unit 131 (S105).

  FIG. 8 is a diagram illustrating a configuration example of a person information storage unit. 8, the person information storage unit 131 stores a person ID, a moving route, a feature amount set, a determination result, and the like for each person who has been detected in the past and has moved out of the imaging region.

  The person ID is identification information for each person, and is assigned, for example, in step S105. In addition, the person ID is information for identifying a person who has appeared in the imaging region of the camera 20 in one information processing device 10. Therefore, the same person ID is not always assigned to a certain person A in each information processing apparatus 10. In addition, the same person ID is not always assigned to the person A that appears at different times.

  The travel route is information that includes a history of position information and time information in a time series. That is, the moving route is information indicating the moving route of the person in the imaging area of the camera 20. Although the example of FIG. 8 shows an example in which the interval is one second, the position information and the time information may be recorded at an even shorter interval.

  The feature amount set is all of a plurality of types of feature amounts extracted for a person. Since the feature amount is extracted for each frame, the result obtained by integrating the feature amount extracted from each frame into one value for each type is stored in the person information storage unit 131. The integration may be performed by selecting or calculating a representative value. For example, a median value may be selected, or an average value may be calculated. Alternatively, a feature amount extracted from a frame in which a person is imaged the largest may be selected.

  The determination result includes, when it is determined that the person is the same person as the person detected in the other information processing apparatus 10, identification information of the other information processing apparatus 10 (hereinafter, referred to as “device ID”). And the person ID of the person in the other information processing apparatus 10 are recorded.

  Subsequently, step S101 and the subsequent steps are repeated.

  FIG. 9 is a flowchart illustrating an example of a processing procedure of a transmission process of the selected feature information. The processing procedure of FIG. 9 is executed in each information processing apparatus 10 in parallel with the processing procedure of FIG.

  In step S201, the selected feature amount determination unit 114 determines whether or not a predetermined time has elapsed from the previous determination time of the selected feature amount. That is, the selected feature amount is determined each time a predetermined time elapses. When a predetermined time has elapsed from the previous determination time (Yes in S201), the selected feature amount determination unit 114 determines whether a person is different based on a feature amount set stored for each person in the person information storage unit 131. The type (selected feature amount) of the feature amount having relatively high suitability is selected (S202). For example, the selected feature amount determination unit 114 acquires a feature amount set of each person related to each record stored in the person information storage unit 131 within a certain time from the current time, and obtains a feature amount for each type of feature amount. Is calculated for the set of. The selected feature amount determination unit 114 selects the types of some feature amounts as selected feature amounts in descending order of variance. This is because the larger the variance is, the more the feature amounts of each person are arranged in the feature space, so that it is considered that the feature amount is more suitable for discriminating each person. However, the selection feature amount may be selected by another method. Further, the number of selected feature amounts may be one or more. A predetermined number of feature values from the top of the variance may be set as the selected feature values. Further, for example, a feature amount having a low correlation with the already selected higher-order feature amount may be preferentially selected. For example, if the color of the upper body and the color of the lower body are common to all the persons, even if the color of the upper body and the color of the lower body are selected as feature quantities, virtually any one of the feature quantities is selected. It can not be expected that the accuracy of determining the difference between persons will be greatly improved as compared with the case where it is performed. Therefore, feature values having low mutual correlation may be selected as the selected feature values. Selection of feature values having low mutual correlation can be realized by mapping all types of feature values to a multidimensional space, performing principal component analysis, and selecting from the first principal component.

  Subsequently, the selected feature transmitting unit 115 transmits the device ID of the information processing device 10 and selected feature information indicating the selected selected feature amount to another information processing device 10 (S203). For example, when the information processing apparatus 10a executes the processing of FIG. 9, the information processing apparatus 10b, the information processing apparatus 10c, and the information processing apparatus 10d are notified of the device ID of the information processing apparatus 10a and the selected feature information. May be transmitted. In the large-scale monitoring system 1 in which a large number of cameras 20 are installed, the selected feature information is transmitted not to all the other information processing apparatuses 10 but to some of the nearby information processing apparatuses 10. May be done. The identification information of the other information processing apparatuses 10 to which the selected feature information is to be transmitted may be stored in each information processing apparatus 10 in advance.

  The selected feature information may be, for example, a bit string in which bits are assigned for each type of feature amount. In this case, the bit corresponding to the type of the feature amount selected as the selected feature amount may be “1”, and the bit corresponding to the type other than the selected feature amount may be “0”.

  In the example of FIG. 9, the selected feature information is notified to the other information processing apparatuses 10 every predetermined period (periodically), but the selected feature information transmitted last time and the selected feature information determined this time are compared. If there is no difference between them, the transmission of the selected feature information need not be performed.

  As described above, the selected feature amount is updated as time passes. Therefore, even when the situation of the imaging region of a certain camera 20 changes according to the time zone, the information processing apparatus 10 uses the feature amount suitable for each time zone to determine whether a person is different. Can be.

  FIG. 10 is a diagram illustrating an example of a change in a situation according to a time zone with respect to an imaging region of the same camera. FIG. 10 shows a situation at a time T1, a time T2, and a time T3 with respect to an imaging region of a certain camera 20.

  At time T1, a situation has occurred in which many people are facing the camera 20. At time T2, a situation occurs in which a small number of people are turning their backs on the camera 20. At time T3, a situation has occurred in which a small number of persons are facing sideways with respect to the camera 20, and a large number of packages are placed between the camera 20 and the persons.

  For example, at time T1, the accuracy (reliability) of the feature amount related to the face of the person is high, but the accuracy of the feature amount related to the color of the upper body and the color of the lower body is considered to be low because the bodies of many people overlap. . Also, at time T2, the accuracy of the feature amount regarding the face of the person is low, but the accuracy of the feature amount regarding the color of the upper body and the color of the lower body is considered to be high. Further, at time T3, the accuracy of the feature amount regarding the face of the person is moderate, the accuracy regarding the color of the upper body is high, and the accuracy regarding the color of the lower body is low.

  In view of the above circumstances, it is conceivable that the feature amount suitable for the selected feature amount differs depending on the time zone even for the same camera 20.

  FIG. 11 is a flowchart illustrating an example of a processing procedure of a reception process of the selected feature information. The processing procedure of FIG. 11 is executed in each information processing apparatus 10 in parallel with the processing procedures of FIG. 7 and FIG.

  The selected feature receiving unit 116 waits for reception of selected feature information transmitted from another information processing device 10 (S301). Upon receiving the device ID and the selected feature information transmitted in step S203 (FIG. 9) executed by one of the other information processing devices 10 (Yes in S301), the selected feature receiving unit 116 transmits the selected feature information to The selected feature is stored in the selected feature storage unit 132 in association with the device ID (S302). If the selected feature information associated with the device ID has already been stored, the existing selected feature information is updated (overwritten) with the newly received selected feature information.

  FIG. 12 is a diagram illustrating a configuration example of the selected feature storage unit. As shown in FIG. 12, the selected feature storage unit 132 stores a bit string as selected feature information in association with the device ID.

FIG. 13 is a flowchart illustrating an example of a processing procedure of a transmission process of a determination request of the same person. The processing procedure of FIG. 13 is executed in each information processing apparatus 10 in parallel with the processing procedures of FIG. 7, FIG. 9, and FIG.
The feature amount acquisition unit 117 is waiting for registration of a new record (person) in the person information storage unit 131 (S401). When a new record (hereinafter, referred to as a “newly registered record”) is registered in the person information storage unit 131 (FIG. 8) by executing step S105 in FIG. 7 (Yes in S401), the feature amount is obtained. The unit 117 acquires the feature data stored in the record (S402).

  Subsequently, the determination request transmission unit 118 selects one information processing device 10 among the other information processing devices 10 that are candidates for the transmission destination of the determination request of the same person (S403). Hereinafter, the selected information processing device 10 is referred to as a “target device”. The list of candidates for the transmission destination of the same person determination request may be the same as the list of other information processing apparatuses 10 to which the selected feature information is transmitted in step S203 in FIG.

  Subsequently, the determination request transmission unit 118 transmits a determination request of the same person including the feature amount acquired in step S402 to the target device (S404). At this time, the feature amount included in the determination request is limited to the feature amount having a bit value of 1 in the selected feature information stored in the selected feature storage unit 132 in association with the device ID of the target device. . That is, the feature amount corresponding to the selected feature amount selected in the target device is transmitted to the target device. It should be noted that there may be a case where the selected feature amount of the information processing apparatus 10 does not match the selected feature amount of the target device. Therefore, the feature amount that is the selected feature amount of the target device and that is the selected feature amount of the information processing device 10 may be transmitted to the target device.

  Subsequently, the determination result receiving unit 122 receives a determination result by the same person determination process performed on the target device (S405). The details of the same person determination process will be described later. When the received determination result indicates that the same person exists (Yes in S406), the determination result includes the person ID of the person determined to be the same person in the target device. Therefore, the determination result receiving unit 122 stores the device ID of the target device and the person ID included in the determination result in the “determination result” of the newly registered record (S407).

  On the other hand, when the determination result indicates that the same person does not exist (No in S406), the determination request transmission unit 118 transmits a determination request regarding the person related to the new record among the candidates for the destination of the determination request of the same person. It is determined whether there is a candidate that has not been transmitted (S408). If there is a corresponding candidate (Yes in S408), steps S403 and subsequent steps are repeated for the information processing apparatus 10 related to the candidate. When there is no corresponding candidate (No in S408), the process returns to step S401. That is, in this case, the search for the same person as the person related to the newly registered record is abandoned.

  In the process of FIG. 13, when it is determined that the same person exists by one of the candidates of the transmission destination of the same person determination request, the search for the same person related to the new record ends. Therefore, the processing of FIG. 13 is executed in a chain in each information processing apparatus 10, and by collecting information stored in the personal information storage unit 131 of each information processing apparatus 10, the movement trajectory of a certain person can be determined. Can be derived.

  The order in which the target devices are selected in step S403 (the order in which the determination requests are transmitted) may be determined in advance according to, for example, the distance of the installation location of the camera 20, or may be dynamically changed according to some condition. You may do so. Alternatively, when the order of the information processing apparatuses 10 is expressed in a ring shape, the selection order in each of the information processing apparatuses 10 is determined such that the information processing apparatus 10 next to the information processing apparatus 10 in the order comes first. You may. In this embodiment, when the information processing apparatuses 10a, b, c, and d are arranged in this order, the selection order in the information processing apparatus 10a is the information processing apparatuses 10b, c, and d. The order may be determined so as to be the information processing apparatuses 10c, d, and a.

  FIG. 14 is a flowchart illustrating an example of a processing procedure of the same person determination process. The processing procedure of FIG. 14 is executed in each information processing apparatus 10 in parallel with the processing procedures of FIG. 7, FIG. 9, FIG. 11, and FIG.

  The determination request receiving unit 119 waits for reception of a determination request for the same person from another information processing apparatus 10 (S501). When the determination request transmitted in step S404 of FIG. 13 executed by any of the other information processing apparatuses 10 is received by the determination request receiving unit 119 (Yes in S501), the same person determination unit 120 performs the determination. It is determined whether or not the same person as the requesting person exists (S502). Specifically, the selected feature is associated with the feature amount included in the determination request and the device ID included in the determination request from the feature amount set stored in the person information storage unit 131. The presence / absence of the same person is determined by collating the selected feature indicated by the selected feature information stored in the storage unit 132. That is, if a record including a feature amount similar to the feature amount included in the determination request is stored in the person information storage unit 131, the person related to the record is the same person as the person related to the determination request. Is determined.

Whether or not the compared feature amounts are similar may be determined by comparing the similarity of the compared feature amounts with a threshold. For example, if the feature to be compared is color information, if one of the feature is (R ₁ , G ₁ , B ₁ ) and the other is (R ₂ , G ₂ , B ₂ ) The following correlation value or the distance between two points in a three-dimensional space of RGB may be calculated as the similarity.

  In this case, if the correlation value or the distance between the two points is equal to or less than the threshold value, the compared feature amounts may be determined to be similar. The similarity may be calculated for the feature amounts of other types by a method suitable for each type.

  The determination as to whether or not they are the same person may be made on condition that the feature amounts of all types compared are similar. Alternatively, it may be determined whether or not the same person is the same person by comparing the sum of the similarities of various types compared with each other and the threshold.

  Subsequently, the determination result transmission unit 121 transmits the determination result by the same person determination unit 120 to another information processing device 10 that has transmitted the determination request (S503). For example, when it is determined that the same person exists, the person ID of the person related to the same person is transmitted. When it is determined that the same person does not exist, information indicating that the same person does not exist (for example, “−1”) is transmitted.

  Next, processing executed in the information processing apparatuses 10a and 10b when the person H moves from the camera 20a to the camera 20b will be specifically described.

  First, in each information processing apparatus 10, the processing procedure shown in FIGS. 9 and 11 is periodically executed. For example, the information processing device 10a selects a selected feature amount based on a feature amount set of each record stored in the person information storage unit 131 of the information processing device 10a. Here, it is assumed that four types of feature amounts, that is, upper body color, lower body color, hair color, and shoe color are included in the feature amount set. For example, when the hair color and the color of the lower body are selected, the information processing device 10b transmits selection feature information indicating that the hair color and the color of the lower body are the selected feature amounts to the information processing device 10b. For example, when each bit of the bit string of the selected feature information corresponds to the order of the upper body color, the lower body color, the hair color, and the shoe color, the bit string {0, 1, 1, 0} is stored in the information processing apparatus 10b. Sent to.

  When the person H appears in the imaging area of the camera 20a, the processing procedure illustrated in FIG. 7 is executed by the information processing device 10a. As a result, a record regarding the person H is stored in the person information storage unit 131 of the information processing device 10a. Here, it is assumed that “005” is assigned as the person ID for the person H.

  In response to the record regarding the person H being stored in the person information storage unit 131, the information processing device 10a executes the processing procedure illustrated in FIG. Here, a request for determining the presence or absence of the same person as the person H is transmitted to the information processing device 10b. However, since the person H has not reached the imaging area of the camera 20b, the information processing device 10b outputs Information indicating that the same person does not exist is returned.

  Thereafter, when the person H appears in the imaging area of the camera 20b, the processing procedure illustrated in FIG. 7 is executed by the information processing device 10b. As a result, a record regarding the person H is stored in the person information storage unit 131 of the information processing device 10b. It is assumed that, in the information processing device 10b, “001” is assigned as the person ID for the person H.

  In response to the record regarding the person H being stored in the person information storage unit 131, the information processing device 10b executes the processing procedure illustrated in FIG. Here, a request for the presence / absence of the same person including the hair color of the record and the color of the lower body is transmitted to the information processing apparatus 10a.

  The information processing device 10a executes the processing procedure illustrated in FIG. 14 in response to the determination request. As a result, the person with the person ID “005” is determined to be the same person as the person requesting the determination. Therefore, the information processing device 10a transmits a determination result including the person ID “005” to the information processing device 10b.

  Upon receiving the determination result, the information processing apparatus 10b compares the device ID of the information processing apparatus 10a with the person ID “005” in the person information storage unit 131 with respect to the “determination result” of the record whose person ID is “001”. Is stored. As a result, it is stored that the person with the person ID “001” in the information processing device 10b and the person with the person ID “005” in the information processing device 10a are the same person.

  As described above, according to the present embodiment, for the information processing apparatus 10 that performs the same person determination processing, a feature amount having relatively high suitability for determining a person is transmitted to the information processing apparatus 10. You. As a result, the amount of data transmitted to derive the trajectory of the person based on the images captured by the cameras 20 installed at a plurality of locations is reduced as compared with the case where all the feature amounts are transmitted. be able to.

  Note that, in the present embodiment, an example has been described in which the subject (object) to be determined to be different is a person, but the difference may be determined for subjects other than the person.

  Note that, in the present embodiment, the feature amount extraction unit 112 is an example of an extraction unit. The selection feature amount determination unit 114 is an example of a selection unit. The selected feature transmitting unit 115 is an example of a transmitting unit. The determination request receiving unit 119 is an example of a receiving unit. The same person determination unit 120 is an example of a determination unit. The camera 20 is an example of an imaging device.

  As mentioned above, although the Example of this invention was described in full detail, this invention is not limited to such a specific embodiment, A various deformation | transformation is carried out within the range of the gist of this invention described in the claim.・ Change is possible.

With respect to the above description, the following items are further disclosed.
(Appendix 1)
An information processing device arranged for each imaging device installed at a plurality of locations,
For a first subject detected from an image captured by a first imaging device corresponding to the information processing device, a plurality of types of feature amounts are extracted,
For each of the plurality of types of feature amounts extracted in the extracting process, the suitability for determining whether the subject imaged by the first imaging device is different is determined, and a part where the suitability is relatively high is determined. Select the feature amount of
Transmitting information for identifying a part of the feature amount selected in the selecting process to another information processing apparatus,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted in the transmitting process is extracted. Receive,
Determining the difference between the first subject and the second subject based on the feature value extracted in the extracting process and the feature value received in the receiving process;
An information processing method comprising performing processing.
(Appendix 2)
The extracting process includes extracting a plurality of types of feature amounts for each of a plurality of subjects detected from an image captured by the first imaging device,
In the selecting process, a variance is calculated for each of the plurality of types of feature amounts extracted in the extracting process, and based on the calculated variance, a part of the feature amount having relatively high suitability is selected. ,
The information processing method according to claim 1, wherein
(Appendix 3)
The process of selecting, for each predetermined period, selecting the partial feature amount,
The transmitting process includes transmitting information identifying the partial feature amount to another information processing device each time the partial feature amount is selected in the selecting process.
3. The information processing method according to claim 1 or 2, wherein
(Appendix 4)
An information processing device arranged for each imaging device installed in a plurality of places,
An extraction unit configured to extract a plurality of types of feature amounts for a first subject detected from an image captured by a first imaging device corresponding to the information processing device;
For each of the plurality of types of feature amounts extracted by the extraction unit, the suitability for determining the difference between the subjects imaged by the first imaging device is determined, and a part of the suitability is relatively high. A selection unit for selecting a feature amount;
A transmitting unit that transmits information for identifying a part of the feature amount selected by the selecting unit to another information processing device,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted by the transmission unit is received. A receiving unit that performs
A determination unit configured to determine a difference between the first subject and the second subject based on the feature amount extracted by the extraction unit and the feature amount received by the reception unit;
An information processing apparatus comprising:
(Appendix 5)
The extraction unit extracts a plurality of types of feature amounts for each of a plurality of subjects detected from an image captured by the first imaging device,
The selection unit calculates a variance for each of the plurality of types of feature amounts extracted by the extraction unit, and based on the calculated variance, selects a part of the suitability where the suitability is relatively high,
4. The information processing apparatus according to claim 4, wherein
(Appendix 6)
The selecting unit selects the part of the feature amounts for each predetermined period,
The transmitting unit transmits information identifying the partial feature amount to another information processing device each time the partial feature amount is selected by the selecting unit.
The information processing apparatus according to claim 4 or 5, wherein
(Appendix 7)
In the information processing device arranged for each imaging device installed in a plurality of places,
For a first subject detected from an image captured by a first imaging device corresponding to the information processing device, a plurality of types of feature amounts are extracted,
For each of the plurality of types of feature amounts extracted in the extracting process, the suitability for determining whether the subject imaged by the first imaging device is different is determined, and a part where the suitability is relatively high is determined. Select the feature amount of
Transmitting information for identifying a part of the feature amount selected in the selecting process to another information processing apparatus,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted in the transmitting process is extracted. Receive,
Determining the difference between the first subject and the second subject based on the feature value extracted in the extracting process and the feature value received in the receiving process;
A program for executing a process.
(Appendix 8)
The extracting process includes extracting a plurality of types of feature amounts for each of a plurality of subjects detected from an image captured by the first imaging device,
In the selecting process, a variance is calculated for each of the plurality of types of feature amounts extracted in the extracting process, and based on the calculated variance, a part of the feature amount having relatively high suitability is selected. ,
The program according to supplementary note 7, characterized in that:
(Appendix 9)
The process of selecting, for each predetermined period, selecting the partial feature amount,
The transmitting process includes transmitting information identifying the partial feature amount to another information processing device each time the partial feature amount is selected in the selecting process.
3. The program according to claim 1 or 2, wherein

Reference Signs List 1 monitoring system 10 information processing device 20 camera 100 drive device 101 recording medium 102 auxiliary storage device 103 memory device 104 CPU
105 Interface device 111 Person detection unit 112 Feature extraction unit 113 Person tracking unit 114 Selected feature determination unit 115 Selected feature transmission unit 116 Selected feature reception unit 117 Feature acquisition unit 118 Judgment request transmission unit 119 Judgment request reception unit 120 Same person Judgment unit 121 Judgment result transmission unit 122 Judgment result reception unit 131 Person information storage unit 132 Selected feature storage unit B Bus

Claims (5)

An information processing device arranged for each imaging device installed at a plurality of locations,
For a first subject detected from an image captured by a first imaging device corresponding to the information processing device, a plurality of types of feature amounts are extracted,
For each of the plurality of types of feature amounts extracted in the extracting process, the suitability for determining whether the subject imaged by the first imaging device is different is determined, and a part where the suitability is relatively high is determined. Select the feature amount of
Transmitting information for identifying a part of the feature amount selected in the selecting process to another information processing apparatus,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted in the transmitting process is extracted. Receive,
Determining a difference between the first subject and the second subject based on a part of the feature amounts selected in the selecting process and the feature amounts received in the receiving process;
An information processing method comprising performing processing. The extracting process includes extracting a plurality of types of feature amounts for each of a plurality of subjects detected from an image captured by the first imaging device,
In the selecting process, a variance is calculated for each of the plurality of types of feature amounts extracted in the extracting process, and based on the calculated variance, a part of the feature amount having relatively high suitability is selected. ,
2. The information processing method according to claim 1, wherein: The process of selecting, for each predetermined period, selecting the partial feature amount,
The transmitting process includes transmitting information identifying the partial feature amount to another information processing device each time the partial feature amount is selected in the selecting process.
3. The information processing method according to claim 1, wherein: An information processing device arranged for each imaging device installed in a plurality of places,
An extraction unit configured to extract a plurality of types of feature amounts for a first subject detected from an image captured by a first imaging device corresponding to the information processing device;
For each of the plurality of types of feature amounts extracted by the extraction unit, the suitability for determining the difference between the subjects imaged by the first imaging device is determined, and a part of the suitability is relatively high. A selection unit for selecting a feature amount;
A transmitting unit that transmits information for identifying a part of the feature amount selected by the selecting unit to another information processing device,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted by the transmission unit is received. A receiving unit that performs
A determination unit configured to determine the difference between the first subject and the second subject based on a part of the feature amounts selected by the selection unit and the feature amounts received by the reception unit;
An information processing apparatus comprising: In the information processing device arranged for each imaging device installed in a plurality of places,
For a first subject detected from an image captured by a first imaging device corresponding to the information processing device, a plurality of types of feature amounts are extracted,
For each of the plurality of types of feature amounts extracted in the extracting process, the suitability for determining whether the subject imaged by the first imaging device is different is determined, and a part where the suitability is relatively high is determined. Select the feature amount of
Transmitting information for identifying a part of the feature amount selected in the selecting process to another information processing apparatus,
Among the feature amounts extracted for the second subject detected from the image captured by the second imaging device corresponding to the other information processing device, the feature amount indicated by the information transmitted in the transmitting process is extracted. Receive,
Determining a difference between the first subject and the second subject based on a part of the feature amounts selected in the selecting process and the feature amounts received in the receiving process;
A program for executing a process.

Images