JP2006067139A - Method, device, and program for searching videos of a plurality of cameras - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To allow moving object detection scenes to be searched from videos of a plurality of cameras so that a user can easily browse them. <P>SOLUTION: A condition input means 28 prescribes a plurality of cameras 11 and a sequence, and an event detection means 24 searches videos of the plurality of cameras recorded in a recording means 21, in accordance with the prescribed cameras and sequence to detect moving object detection scenes as events. In the case of reproduction, a search means 22 searches a moving object detection scene in accordance with the prescribed cameras and sequence, and a reproducing means 23 displays the moving object detection scene in a video reproduction picture 132 on a display device, and a map display means 26 displays a map picture 131 which is displayed by moving a reproduction cursor synchronized with a reproduction time of the moving object detection scene, on the display device 13 in accordance with a sequence of cameras linked by arrows on the map picture 131. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a multi-camera video search device, a multi-camera video search method, and a multi-camera video search program for searching for a moving object detection scene or the like from a plurality of camera videos.

  In the video monitoring field, motion detection is used when performing abnormality detection, intruder detection, and the like. In the moving object detection, a difference image is generated from two different frame images, and it is determined that a moving object has appeared when a change in luminance of a certain level or more is confirmed. This moving object detection function is effective not only for the purpose of notifying the user of an alarm indicating an abnormality, but also for improving the efficiency of storage such as storing only the video portion detected by the moving object.

Therefore, as a conventional multi-camera video search device, for example, one or a predetermined number of videos from a plurality of monitoring cameras, a diagram showing an installation area of the monitoring camera corresponding to the videos, and a predetermined command for performing a monitoring operation Are displayed together on the monitor and a predetermined command on this monitor is selected, and a predetermined operation including switching of the image display state from the monitoring camera is performed, and anyone can display the image from the monitoring camera. There is an apparatus that can easily perform a monitoring operation including a state change (see, for example, Patent Document 1).
Japanese Patent No. 2791411

  However, when viewing images from multiple cameras at the same time on multiple screens, it is difficult for the user to grasp which position the moving object will appear in the image of the camera installed next, and at what timing the moving object will appear. there were. Therefore, it is necessary to wait for the appearance of a moving object over a plurality of videos, and there is a problem that it is not possible to concentrate on browsing.

  Even in the conventional multi-camera video search device described in Patent Document 1, the camera icon of the map corresponding to the video being played is colored. It does not consider the presence or absence, and the above-mentioned problem cannot be solved.

  Therefore, the present invention provides a multi-camera video search device, a multi-camera video search method, and a multi-camera video search program that can search a motion detection scene from videos of a plurality of cameras so that a user can easily browse. Objective.

  In order to solve the above-described problem, a multi-camera video search device according to the present invention is a multi-camera video search device that searches video of a plurality of cameras, and a condition input unit that specifies the plurality of cameras and their order. And a moving object detection scene search means for searching for a moving object detection scene in which a moving object is detected according to the specified plurality of cameras and the order from the plurality of camera images.

  Thereby, since the moving body detection scene which detected the moving body according to the designated some camera and order is searched, a moving body detection scene can be searched so that a user may browse easily.

  The multi-camera video search device according to the present invention may further include a map display unit configured to display a map on which camera icons indicating positions of the plurality of cameras are displayed, and the condition input unit includes the map The above camera icons are connected with arrows to designate the plurality of cameras and their order.

  As a result, the camera icons on the map are connected with arrows to specify the plurality of cameras and their order, so that the user can input search conditions by an intuitive operation based on the camera icons on the map. It becomes.

  In the multiple camera video search device according to the present invention, in the configuration, the map display unit calculates a connection order of camera icons to be searched each time a plurality of camera icons are connected by arrows, A configuration is adopted in which a numerical value is displayed next to.

  Thereby, the user can grasp the order relation of a plurality of cameras by the numerical value displayed near each camera icon without following the arrow.

  Further, in the multi-camera video search device of the present invention, in the above configuration, when the motion detection scene searched by the search means is played back, a camera icon that captures the motion detection scene being played back is highlighted. Take.

  Thereby, the user can grasp at a glance the camera detecting the moving object.

  In the multi-camera image search device according to the present invention, in the above configuration, the map display means captures a camera icon that captures the motion detection scene being played back when the motion detection scene searched by the search means is played back. The imaging field-of-view range is superposed and displayed on the map.

  Thereby, the user can grasp the approximate spatial position of the moving object being detected.

  In the multiple camera video search device of the present invention, in the configuration, the map display means sets a playback cursor indicating a playback time when the motion detection scene searched by the search means is played back, and the condition input means. A configuration is adopted in which the image is moved and displayed on the arrow input by the camera toward the camera icon in which a moving object appears.

  Accordingly, the user can grasp the appearance of the moving object such as the position at which the moving object appears next and the timing at which the moving object appears.

  In the multiple camera video search device of the present invention, in the configuration, the map display means sets a playback cursor indicating a playback time when the motion detection scene searched by the search means is played back, and the condition input means. The moving object detection scene is displayed while being moved toward the camera icon in which a moving object newly appears on the arrow input by the button, and the moving object detection scene is reproduced by the reproduction cursor.

  Thus, on one screen, the user can view the moving object detection scene while grasping at which position the moving object appears next, and at which timing the moving object appears.

  In the multi-camera video search device according to the present invention, in the configuration, the map display unit can set a playback cursor indicating a playback time at an arbitrary position on the arrow input by the condition input unit, and A configuration is adopted in which the searched motion detection scene at the playback time based on the position on the arrow input by the condition input means where the playback cursor is positioned is played back.

  Thus, when a user wants to view a moving object detection scene (video portion) in which a moving object moves across a plurality of cameras, the user designates the position of the reproduction cursor at an arbitrary position by the condition input means, thereby moving the moving object at an arbitrary reproduction time. The detection scene can be reproduced, and the user does not need to estimate and specify the time when the user seems to straddle the cameras.

  In addition, the multi-camera image search method of the present invention employs a method of searching for a moving object detection scene in which a moving object is detected according to a plurality of cameras and order specified from a plurality of camera images. Thereby, a moving body detection scene can be searched so that a user may browse easily.

  Further, the multi-camera video search program of the present invention adopts a configuration in which a computer is executed to search for a moving object detection scene in which a moving object is detected in accordance with a plurality of cameras specified from a plurality of camera videos and the order. Thereby, a moving body detection scene can be searched so that a user may browse easily.

  According to the present invention, since a moving object detection scene in which a moving object is detected in accordance with a plurality of designated cameras and an order is searched from a plurality of camera images, the moving object detection scene can be searched so that the user can easily browse.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the form of a system and an apparatus, but can be implemented in the form of a method, a program for realizing the method, and a storage medium storing the program, as supported by the description of the embodiment. it can.

Embodiment 1 FIG.
FIG. 1 is a system overview of a multi-camera video search device 12 according to Embodiment 1 of the present invention.

  As shown in FIG. 1, the video monitoring system includes a plurality of cameras 11 that capture a monitoring area, a multi-camera video search device 12, and a display 13.

  The camera 11 captures individual monitoring areas and inputs them to the multiple camera video search device 12. The multi-camera video search device 12 displays the video of the camera that captured each monitoring area on the display 13. The mouse 14 is a pointing device for performing a GUI operation related to video search / playback. Since the camera video is composed of various scenes, and the camera 11 of the first embodiment is for monitoring, a moving object detection scene in which a moving object is detected and a scene in which a moving object is not detected It will consist of

  FIG. 2 is a functional block diagram showing the configuration of the multi-camera video search device 12 according to the embodiment of the present invention.

  As shown in FIG. 2, the multi-camera video search device 12 includes a storage unit 21, a search unit 22, a playback unit 23, an event detection unit 24, a video input unit 25, a map display unit 26, and a video output unit. 27 and condition input means 28. Note that the search means 22 and the event detection means 24 of the first embodiment correspond to the moving object detection scene search means of the present invention.

  The storage means 21 is composed of a storage device such as a hard disk or a large-capacity memory, and includes a camera image, map information including a monitoring area, camera icon information, event information indicating an event detection result, and a search condition indicating a search condition. Information, search result information indicating search results, and the like are stored. The camera video is input from the video input means 25. The search condition information is input from the condition input unit 28 by the user, and is used when the search unit 22 searches the video. The search result information is a result of searching the camera video by the search means 22 and is used when the video is played back by the playback means 23. Note that map information and camera icon information are stored in advance.

  The search means 22 searches for a video scene including a motion detection scene having an event pattern designated by the user based on the search condition information and event information stored in the storage means 21. The search result of the camera video is output to the storage means 21 as search result information. An example of event information is shown in FIG. 6 to be described later, an example of search condition information is shown in FIG. 7 to be described later, and an example of search result information is shown in FIG. 8 to be described later.

  The reproduction unit 23 reads the camera video stored in the storage unit 21 and outputs it to the video output unit 27. Also, based on the search result information 214 stored in the storage means 21, the playback status during video playback is calculated. The reproduction status is output to the map display means 26. Note that the playback unit 23 is not essential for the multi-camera video search device 12, and the multi-camera video search device 12 is not provided with the playback unit 23, and the playback unit 23 provided outside the apparatus plays back. You may comprise so that it may do.

  The event detection unit 24 analyzes the camera video acquired from the plurality of cameras 11 via the video input unit 25 and the camera video stored in the storage unit 21, and determines whether a moving object exists in the video, that is, the moving object is detected. The detected moving object detection scene is detected. The moving object detection scene is output to and stored in the storage means 21 as event information.

  The video input means 25 acquires video from the plurality of cameras 11, outputs it to the event detection means 24, confirms the moving object detection scene in the video, and then outputs it to the storage means 21.

  The map display means 26 displays a map of the monitoring area on which the camera icons indicating the positions and orientations of the plurality of cameras 11 stored in the storage means 21 are displayed on the display 13. In addition, the playback status during video playback calculated by the playback means 23 is superimposed on the map.

  The video output unit 27 displays the camera video stored in the storage unit 21 on the display 13.

  The condition input means 28 acquires search conditions and search / playback commands from the user. The search condition is output and stored in the storage means 21 as search condition information. Further, when the search command is acquired, the search unit 22 is notified, and when the playback command is acquired, the playback unit 23 is notified.

  FIG. 3 shows an example of the data structure of the camera video data 110 stored in the storage unit 21. The camera video data 110 is managed in units of frame images. As shown in FIG. 3, the frame image includes information on a frame number, a frame size, a time stamp, and a camera number as header information. These header information is followed by video data.

  FIG. 4 shows an example of the map screen 131 displayed by the map information stored in the storage unit 21. As shown in FIG. 4, the map screen 131 is an image or video of a map 51 showing indoor and outdoor monitoring areas monitored by a plurality of cameras 11. That is, on this map screen 131, camera icons I1 to I5 indicating the positions and orientations of the plurality of cameras 11 are displayed on the map 51 indicating the monitoring area, and the monitoring target location such as a table in the monitoring area is displayed. A monitoring target icon I6 indicating the size and size is displayed.

  FIG. 5 shows an example of the camera icon information 211 stored in the storage unit 21. The camera icon information 211 is information relating to the camera icons I1 to I5 to be superimposed on the map 51 as shown in FIG. 4, and as shown in FIG. 5, for each camera icon I1 to I5, the camera position for each camera number. As information, information on coordinates (X, Y) on the map and camera direction D (shooting direction or angle) is set in association with each other. In the case shown in FIG. 4, N is 1 to 5.

  FIG. 6 shows an example of event information 212 stored in the storage unit 21. The event information 212 is information related to a moving object detection scene that is a scene in which the event detection unit 24 analyzes individual camera images and detects a moving object, and is managed by an event number. In other words, as shown in FIG. 6, in the event information 212, information on the start time and end time of the motion detection scene is associated with each motion detection scene for each event number.

  FIG. 7 shows an example of the search condition information 213 stored in the storage unit 21. The search condition information 213 is input by the user via the condition input means 28, and is information including a condition for searching for a scene in which a moving object is detected with a pattern to be viewed. Specifically, as shown in FIG. 7, the order number and the camera number information are associated with each other. In the case of FIG. 7, the scene in which the moving object is detected in the order of the camera number C1, C2, C3,. Indicates to search. Note that one or a plurality of search condition information 213 can be set according to the user's preference or the like, and the event information 212 shown in FIG. 6 is referred to based on the set one or more search condition information 213. The search means 22 searches for a moving object detection scene.

  FIG. 8 shows an example of the search result information 214 stored in the storage unit 21. The search result information 214 is information related to the time when each camera 11 detects a moving object, that is, the start time of the moving object detection scene, according to the search condition information 213 specified by the user. As shown in FIG. 8, the search result information 214 is set by associating the camera number to be searched with the motion detection time. When a plurality of search condition information 213 are set, a plurality of search result information 214 is set accordingly.

  Next, the screen displayed on the display 13 in this Embodiment is demonstrated.

  FIG. 9 shows an example of the map screen 131 displayed when the application of the multi-camera video search device 12 in the present embodiment is started. When the application of the multi-camera video search device 12 is activated, a map screen 131 as shown in FIG. 4 stored in the storage unit 21 is read as it is and displayed on the display 13. From the map screen 131 shown in FIG. 9, the positions and orientations of the five cameras 11 arranged in the room or the like can be known from the camera icons I1 to I5, and the monitoring object icon I6 can be used in the management target area. The position and size of the monitoring target can be known.

  FIG. 10 shows an example of a map screen 131 displayed after the application of the multi-camera video search device in this embodiment is started. That is, on the map 51 on which the camera icons I1 to I4 indicating the positions and orientations of the five cameras 11 arranged in the room to be monitored displayed as shown in FIG. An arrow 53 and a playback cursor 54 are displayed.

  The arrow 53 is displayed by connecting the camera icons I1 to I4 selected by the user using the mouse 14 by the user. When the arrow 53 is connected by the camera icon, the color changes or is highlighted to distinguish it from the camera icon that is not connected, so as to indicate that the selection is made by the arrow 53. Further, the arrow 53 designates a camera for searching for a moving object detection scene and its order, that is, an event pattern, and a candidate selection tab 55 shown in FIG. 11 is provided for each arrow 53. .

  The playback cursor 54 moves toward the camera icons I1 to I4 in which the moving object appears next during playback of the moving object detection scene, and moves in synchronization with the playback time of the camera video.

  FIG. 11 shows an example of the video playback screen 132 displayed after the application of the multi-camera video search device in the embodiment is started. That is, on the display 13 of the present apparatus, after the application is started, the playback unit 23 causes the playback unit 23 to display a video playback screen 132 displaying a candidate selection tab 55, a video playback window 56, a search button 57, and a playback button 58 as shown in FIG. Is displayed.

  The candidate selection tab 55 is used when selecting a scene including a motion detection scene of an event pattern that the user wants to browse, and corresponds to a combination of a plurality of cameras 11 selected by the user by the arrow 53 and their order. is there. In FIG. 11, four candidate selection tabs 55 are provided, which indicates that the user has designated an event pattern of a combination of a plurality of four cameras 11 and their order.

  Here, the event pattern is, for example, for the user to select a motion detection scene of an event pattern corresponding to the connection order indicated by the arrow 53 in which the camera icons I1 to I5 are connected using the mouse 14. . For example, in the case of FIG. 12 to be described later, since they are connected by the arrow 53 in the order of camera icons I1 → I2 → I3 → I4, as shown in FIG. 13 to be described later, a moving object including events E1 → E4 → E6 → E8 Search the detection scene. The number of candidate selection tabs 55 is based on the motion detection scene of the event pattern selected by the user included in the search target video, that is, the combination of the plurality of cameras 11 selected by the user and their order.

  The video playback window 56 is set for each candidate selection tab 55, that is, is provided for each of the plurality of cameras 11 selected by the user, and displays the video of each camera 11 respectively.

  The search button 57 is a button for a user to issue a search command, and the play button 58 is a button for the user to execute a play command. The playback button 58 serves as a stop button during video playback.

  FIG. 12 shows an example of the map screen 131 displayed when the moving object detection scene is reproduced by the multi-camera video search device according to the present embodiment. That is, when the application is activated, the map 51 indicating the monitoring target such as a room displayed as shown in FIG. 9 and the camera icons I1 to I4 indicating the positions and orientations of the five cameras 11 arranged in the room are displayed. Further, the condition input means 28 draws an arrow 53 by the user to designate a large number of cameras 11 and their order, and a moving object detection scene is detected according to the designated cameras 11 and order, and the video reproduction screen 132 shown in FIG. When the motion detection scene is played back, the map screen 131 displays playback cursors C1 to C3 indicating the playback time of the motion detection scene.

  Specifically, camera icons I1 to I4 corresponding to the designated camera 11 are connected by an arrow 53 according to a user instruction, and the playback cursor 54 is positioned on the arrow 53 in accordance with the playback time of the motion detection scene. It is displayed while moving toward the camera icon where a new moving object appears.

  In FIG. 12, C1 to C3 indicate cases where the playback cursor 54 is between the camera icons I1 and I2, between I2 and I3, and between I3 and I4. Only one playback cursor 54 is displayed.

  FIG. 13 is a diagram illustrating a method of searching for a motion detection scene from the video of each camera 11 selected by the user.

  That is, for example, the user selects four cameras 11 via the condition input unit 28, and the images of the four cameras 11 are denoted by L1 to L4. Assume that each of the videos L1 to L4 includes events E1 to E8 that are moving object detection scenes.

  In addition, since the pattern of the moving object detection scene is each of the videos L1, L2, L3, and L4, when the search unit 22 extracts the event E1 from the video L1, an event that starts after the start time of the event E1 is detected from the video L2. By searching, the event E4 is extracted from the video L2. Similarly, an event E6 starting after the start time of the event E4 of the video L2 is extracted from the video L3, and an event E8 starting after the start time of the event E6 of the video L3 is extracted from the video L4.

  The reproduction cursors C1 to C3 indicate the display timings of the events E1, E4, and E6 in the videos L1 to L4, respectively. The reproduction cursor 54 moves on the map 51 as time passes. It shows that the playback cursor changes from C1 to C3.

  Further, the position of the reproduction cursor 54 can be arbitrarily changed by the user, and an event at a time corresponding to the designated position, that is, a moving object detection scene can be displayed.

  After the application is started, the map screen 131 shown in FIGS. 10 and 12 and the video playback screen 132 shown in FIG. 11 are displayed at the same time, displayed on the same display 13 simultaneously, or on different displays 13. At the same time, a camera image such as a moving object detection scene is observed on the image reproduction screen 132 while referring to the map screen 131.

  Next, the operation will be described with reference to the drawings.

  FIG. 14 is a flowchart showing video storage process processing including event detection according to Embodiment 1 of the present invention.

  First, the event detection means 24 analyzes the video of each camera 11 and detects a moving object (S1100). For the detection of a moving object, for example, a threshold process of luminance is performed on a difference image between two frame images, and if there is a luminance change equal to or greater than the threshold, it is determined that there is an event because the camera image has detected a moving object. On the other hand, when the luminance conversion is equal to or less than the threshold value, it is determined that there is no event because the moving object is not reflected in the camera video. Then, the event detection unit 24 stores the moving object detection scene in which the moving object is detected in the storage unit 21 as the event information 212 (S1200). The event information 212 is managed for each camera 11, and an event start time and an end time are given as shown in FIG. Finally, the storage means 21 stores the camera video (S1300). The moving object detection by the event detection unit 24 may be video data input from the camera 11 through the video input unit 25, or may be stored in the storage unit 21 and targeted for video data. In the moving object detection process of S1100, when a moving object detection window, that is, a place where a moving object is to be detected is set in advance or by a user in a limited manner, the moving object detection process is performed only within the window. . The above is the flowchart of the video storage process.

  FIG. 15 is a flowchart showing the startup process processing of the video search application in the embodiment of the present invention.

  First, the map display means 26 reads the map screen 131 stored in the storage means 21 (S2100) and reads the camera icon information 211 (S2200). Then, after the read map screen 131 is displayed on the display 13 as a map image (S2300), the camera icons I1 to I4 are displayed superimposed on the map image (S2400). The camera icon information 211 includes the position and orientation of the camera icon to be superimposed on the map image. The above is a flowchart of the startup process of the video search application.

  When the map screen 131 as shown in FIG. 9 is displayed on the display 13, the user connects the camera icons I 1 to I 4 with the arrow 53 using the mouse 14 as shown in FIG. The connection order indicated by the arrow 53 is a search condition. The input search condition is stored in the storage unit 21 as search condition information 213. In addition to the user input, the present apparatus may automatically recognize and connect the camera icons I1 to I4 with the arrow 53 or guide the connection on the map screen 131. For example, when the monitoring area is indoors and there are an entrance, a window, and an exit, the camera icons I1 to I4 near the entrance are searched for the camera icons I1 to I4 near the exit and automatically linked by the arrow 53. can do.

  FIG. 16 is a flowchart showing the process of searching for a moving object detection scene in the embodiment of the present invention.

  The search means 22 first confirms whether or not the input of search conditions has been completed by the user or the like (S3000). If the input of the search condition is not completed (S3000 “NO”), an error is displayed (S3100), and the process returns to the search condition input completion confirmation process of S3000.

  On the other hand, when the input of the search condition is completed (S3000 “YES”), the search unit 22 reads the search condition information 213 in which the user or the like has designated the camera 11 and its order from the storage unit 21 (S3200). The event information 212 is read (S3300).

  Next, the search means 22 searches the event information 212 for a motion detection scene of an event pattern based on the search condition (S3400). Here, since there may be a plurality of event detection scenes of event patterns based on the search condition, the process is performed until all the motion detection scenes to be searched are searched.

  If a motion detection scene with an event pattern has been searched (S3500 “YES”), it is determined whether or not the search for all event information 212 has been completed for the motion detection scene with all event patterns (S3600). If the search of all event information 212 has not been completed for the motion detection scenes of all event patterns (S3600 “NO”), the candidate selection tab 55 is added to the motion detection scene of the event pattern (S3700). The process returns to the event pattern moving object detection scene search process of S3400. When a motion detection scene of an event pattern is searched in this way, a candidate selection tab 55 is added for each of the searched motion detection scenes of the event pattern, and the user can select this candidate as shown in FIG. By selecting the selection tab 55, it is possible to select a motion detection scene of an event pattern based on the search condition. The above is the process of searching for a moving object detection scene.

  FIG. 17 is a detailed flowchart of the moving object detection scene search process of S3400 shown in FIG.

  First, the search means 22 determines whether or not it is the first time (S4000). Only in the first time (S4000 “YES”), the reference time is initially set (S4100). The reference time is a parameter for grasping the order in which events are detected when searching for a video such as a moving object detection scene. The initial setting of the reference time is performed only for the first time. Then, in the search condition information 213, the first camera 11 in the order is set as a search target camera (S4200).

  Next, in a plurality of cameras 11 to be searched, an event detected after the reference time, that is, a motion detection scene is searched (S4300).

  When the event search is successful (S4400 “YES”), the start time of the found motion detection scene is set and updated as a new reference time (S4500), and if all the search target camera images have not been completed (S4600). “NO”), the search continues for the next camera image (S4700). In this manner, an event that is a moving object detection scene is searched for up to the last camera in the specified order in the search condition information 213. If no event is found during the event search process, the process ends in S3400.

  For example, in the case shown in FIG. 13, the moving object detection scene is searched for the videos L1 to L4. As a result, in the example of FIG. 13, when the start time of the event E1 is the reference time in the video L1, the motion detection scene detected after the reference time is searched and the reference time is updated, so the video L2 Then, event E4 is a search target event, event E6 is a search target event in video L3, and event E8 is a search target event in video L4. The above is a detailed flowchart of the event pattern moving object detection scene search process of S3400.

  FIG. 18 is a detailed flowchart of the event search process in S4300 of FIG.

  The search means 22 first determines whether or not it is the first time (S5000), and only when it is the first time (S5000 “YES”), initializes the event number (S5100). The event number is an index indicating an individual event. Initial setting is performed only for the first time, and after the second time, the next event number stored in the storage means 21 is read (S5200).

  Next, it is confirmed whether or not a motion detection scene that is an event indicated by the event number exists (S5300). If an event is found (S5400 “YES”), the start time of the motion detection scene is referred to (S5500). Then, the reference time, which is the start time of the motion detection scene in the previous camera image, is compared with the start time of the motion detection scene in the camera image (S5600), and the start time is the reference. If it is later than the time, it is determined that the order relationship has been established (S5700 “YES”), the search result information 214 is stored in the storage means 21 (S5800), and the event number of the moving object detection scene is stored in the storage means 21. Store (S5900). In S5600, when comparing the magnitude relation between the reference time of the motion detection scene in the camera image one order before and the start time of the motion detection scene of the camera image, the size of the time difference is also taken into consideration. Anyway. In other words, even if the start time of the motion detection scene of the camera image is later than the reference time, if it is too late in time, it is considered to be a less relevant motion detection scene. Also good.

  Here, the event number is stored in step S5900 in this way because the apparatus performs synchronized playback of the images of a plurality of cameras. Therefore, when the event search is performed next time, the event number previously searched is stored. Since the search process can be performed from the event that is later in time, the event number is set so that the event number that is earlier in time than the previously searched event number, that is, the event number does not return to the previous event. It is something to remember. In this way, when there are a plurality of events and the event search process is performed many times, there is no need to return to the previous event in time, so the efficiency of the event search process is greatly improved. Of course, the process of storing the event number of the moving object detection scene in S5900 in the storage unit 21 may be omitted.

  On the other hand, when the start time is earlier than the reference time (S5700 “NO”), it is determined that the order relationship has not been established, the process moves to the next event, that is, the motion detection scene (S5750), and the process returns to S5300. Search for events that satisfy the condition. The above is the detailed flowchart of the event search process of S4300.

  FIG. 19 is a flowchart showing video playback process processing according to the embodiment of the present invention.

  The reproduction means 23 first confirms whether the event search has been completed (S6000). If the event search is not completed (S6000 “NO”), an error is displayed (S6100). When the search is completed (S6000 “YES”), the event search result information 214 stored in the storage means 21 is read (S6200), and the moving object detection scene that is a frame image corresponding to the inspection result information is read. Read (S6300). Then, the time stamp included in the header of the frame image is referred to (S6400), and the position of the playback cursor 54 is calculated (S6500).

  The position of the playback cursor 54 starts from obtaining two camera numbers that sandwich the playback cursor 54. This can be grasped from the relationship between the time stamp of the frame image and the event detection time with reference to the search result information 214 shown in FIG.

  For example, in FIG. 8, when the time stamp T has a relationship of detection time DT2 <T <DT3, the camera icons I1 and I2 corresponding to the camera numbers 1 and 2 sandwich the playback cursor 54.

  Then, as shown in FIG. 20, the coordinates of the camera icon I1 with the camera number 1 are (X1, Y1), the time when the event is detected is DT1, the coordinates of the camera icon I2 with the camera number 2 (X2, Y2), and the event is detected. Assuming that the time is DT2 and the time stamp is T, the coordinates (XC, YC) of the playback cursor 54 are given by the following (Expression 1) and (Expression 2).

XC = (X2−X1) × (T−DT1) / (DT2−DT1) + X1 (Formula 1)
YC = (Y2−Y1) × (T−DT1) / (DT2−DT1) + Y1 (Formula 2)

  Finally, the map display means 26 displays the playback cursor 54 on the map screen 131 at the position of the calculated coordinates (XC, YC) (S6600), and the playback means 23 displays a frame corresponding to the time of the playback cursor 54. The moving object detection scene which is an image is displayed (S6700). The display of the frame image is repeated by moving to the next frame image until a stop command is issued from the user (S6800). The above is the video playback process.

  FIG. 21 is a flowchart showing video playback process processing using the playback cursor 54 in the embodiment of the present invention.

  The reproduction means 23 first confirms whether or not a plurality of cameras to be viewed and a candidate selection tab 55 (see FIG. 11) indicating the order are selected, and selection of the plurality of cameras to be viewed and their order has been completed (see FIG. 11). S7000). If the selection of the candidate selection tab 55 has not been completed (S7000 “NO”), an error is displayed (S7100). If the selection has been completed (S7000 “YES”), the position of the playback cursor 54 is acquired (S7200) and converted into a time stamp format (S7300).

  Conversion to the time stamp format starts from obtaining the coordinates of two camera icons sandwiching the playback cursor 54 and the time at which the event was detected. In the example of FIG. 20, two camera icons I1 and I2 sandwiching the playback cursor 54 have the camera icon I1 with the camera number 1 whose coordinates are (X1, Y1), the time DT1 when the event is detected, and the coordinates are (X2, Y2) is the camera icon I2 of camera number 2. The coordinates of the two camera icons I1 and I2 are (X1, Y1) and (X2, Y2), respectively, the times when the event is detected are DT1 and DT2, respectively, and the coordinates of the playback cursor 54 are (XC, YC). ), The time stamp T is given by (Expression 3) and (Expression 4) below.

T = (T2−DT1) × (XC−X1) / (X2−X1) + DT1 (Formula 3)
Or
T = (T2−DT1) × (YC−Y1) / (Y2−Y1) + DT1 (Formula 4)

  Then, the frame image of the motion detection scene having the time stamp closest to the calculated time stamp T is read from the storage means 21 (S7400), and the video playback window 56 of the video playback screen 132 shown in FIG. Is displayed inside (S7500). The video playback process using the playback cursor 54 has been described above.

  As described above, according to the first embodiment, a moving object is detected from a plurality of camera images according to the plurality of cameras 11 and the order designated by the arrow 53 on the map screen 131, and the moving object detection scene is displayed on the video reproduction screen 132. Therefore, it is possible to designate a plurality of cameras 11 and the order to be easily searched on the map screen 131, and to reproduce the motion detection scene searched according to the designation on the video reproduction screen 132. The moving object detection scene can be searched and reproduced so that the user can easily browse.

  As a result, if the user displays the map screen 131 and the video playback screen 132 side by side on the display 13 and browses them, the moving object will appear in the camera 11 installed next, and at what timing the moving object will appear. It is possible to grasp the appearance of moving objects such as whether they appear.

  In the first embodiment, since the camera icons I1 to I4 on the map screen 51 are connected by the arrow 53 to designate the plurality of cameras 11 and their order, the user can set the camera icons I1 to I4 on the map 51. Based on this, it becomes possible to input search conditions by an intuitive operation.

  In the first embodiment, as shown in FIG. 1 and FIG. 2, the multi-camera video search device 12 is provided with the storage means 21 and the playback means 23. However, in the present invention, For example, the storage means 21, the playback means 23, the video input means 25, and the video output means 27 are not provided in the multi-camera video search device 12, and the same as the plurality of cameras 11 and the display 13 is provided. An external video storage device is connected to an externally connected storage means 21, playback means 23, video input means 25, video output means 27, or a video storage device or video playback device having only a part thereof. Of course, it is possible to read a camera video from a camera, search for a motion detection scene, and play it back on a video playback device or the like connected to the outside.

Embodiment 2. FIG.
The second embodiment shows another example of the map screen 131 displayed on the display 13.

  FIG. 22 shows an example of the map screen 131 displayed on the display 13 of the second embodiment.

  In the second embodiment, as shown in FIG. 22, each time the user connects the camera icons I1 to I4 to be searched with arrows, the connection order of the camera icons I1 to I4 is dynamically calculated, and the camera icon I1 is calculated. Next to ~ I4, a connection order icon 81 displayed as a numerical value such as “1”, “2”, “3”, “4” is displayed. Thereby, the user can grasp the order relationship of the cameras 11 by numerical values without following the arrow.

Embodiment 3 FIG.
The third embodiment shows another example of the map screen 131 displayed on the display 13.

  FIG. 23 shows an example of the map screen 131 displayed on the display 13 of the third embodiment.

  In the third embodiment, as shown in FIG. 23, when the camera image is reproduced, the camera icon I3 of the camera 11 that is detecting the moving object is highlighted. Thereby, the user can grasp at a glance the camera detecting the moving object.

Embodiment 4 FIG.
The fourth embodiment shows another example of the map screen 131 displayed on the display 13.

  FIG. 24 shows an example of the map screen 131 displayed on the display 13 of the fourth embodiment.

  In the fourth embodiment, as shown in FIG. 24, when a camera image is played back, a field-of-view icon 83 indicating the imaging field-of-view range of the camera icon I3 of the camera 11 that is moving is detected and superimposed on the map 51. Thereby, the user can grasp the approximate spatial position of the moving object being detected.

Embodiment 5. FIG.
The fifth embodiment shows another example of the map screen 131 displayed on the display 13.

  FIG. 25 shows an example of the map screen 131 displayed on the display 13 of the fifth embodiment.

  In the fifth embodiment, as shown in FIG. 25, when a camera image is reproduced, a large reproduction cursor 54 is configured by a window capable of reproducing an image toward a camera icon in which a moving object newly appears, and the reproduction is performed. The cursor 54 is moved while reproducing the camera image, in particular, the moving object detection scene searched by the search means 22.

  Thereby, on one screen, the user can grasp at which position the moving object appears next, and at which timing the moving object appears, and browse the actual camera image. Is possible. The playback cursor 54 may sequentially display the motion detection scenes according to the camera 11 specified by the user and the order thereof, but the position may be freely or arbitrarily settable by the user.

  As a result, on one screen, the user can grasp at which position the moving object appears next, and at which timing the moving object appears, and can view the actual camera image. It becomes possible.

  Further, since the position of the playback cursor 54 can be freely or arbitrarily set by the user, it is possible to easily search for and reproduce the moving object detection scene in the camera 11 or location that the user wants to play back.

Embodiment 6 FIG.
In the above-described embodiment, the multi-camera video search device 12 according to the present invention has been functionally configured as shown in FIG. 2, but in the fifth embodiment, the multi-camera image search device 12 is configured by a general-purpose personal computer or the like. Show the case.

  FIG. 26 shows a configuration of a multi-camera video search apparatus in Embodiment 6 of the present invention.

  The multi-camera video search device 12 is composed of a general-purpose computer. As shown in FIG. 26, the multi-camera image search device 12 includes a CPU 31, a main storage device 32, a secondary storage device 33 such as a hard disk, a capture board 34, and the like. And a video board 35.

  The capture board 34 serves as an interface for capturing a large number of images from the camera 11 into the camera image search device 12.

  Similar to the storage unit 21 shown in FIG. 2, the main storage device 32 includes a camera image and a map screen 131 including a monitoring area, camera icon information 211, event information 212 indicating an event detection result, and a search condition indicating a search condition. Information 213 and search result information 214 indicating search results are stored. The map screen 131 and the camera icon information 211 are stored in advance.

  In the case of the sixth embodiment, the multi-camera video search program is read from the storage medium and installed in the secondary storage device 33, or is downloaded and installed via the network. By executing the multi-camera video search program, the functions of the respective units of the multi-camera video search device 12 shown in FIG. 2 are achieved, and the same effects as those of the first embodiment can be obtained.

  The multi-camera video search device, multi-camera video search method, and multi-camera video search program of the present invention search for a moving object detection scene in which a moving object is detected in accordance with a plurality of designated cameras and order, so that the user can easily browse. In this way, it is possible to search for a moving object detection scene, and a plurality of camera images include images by combinations of various sensors, and a surveillance camera system in a room or store using a plurality of cameras, Storage media such as video tapes and DVDs, or storage devices such as hard disk recorders, which store a large number of camera images in wide-area parking lots, surveillance camera systems such as station premises, airfields, streets, and roads, and storage devices such as these Read video data from media, computer or TV receiver with those storage devices, video editing置等, multiple cameras video searching apparatus for searching the images captured by the plurality of cameras, multiple cameras video searching method, and multiple cameras image search program, and further is useful in a storage medium or the like storing the program.

The figure which shows the system outline | summary of the multi-camera video | video search apparatus which concerns on Embodiment 1. FIG. Functional block diagram showing the configuration of a multi-camera video search apparatus according to Embodiment 1 The figure which shows an example of the data structure of the camera image | video memorize | stored in a memory | storage means The figure which shows an example of the map screen memorize | stored in the memory | storage means The figure which shows an example of the camera icon information memorize | stored in the memory | storage means The figure which shows an example of the event information memorize | stored in a memory | storage means The figure which shows an example of the search condition information memorize | stored in a memory | storage means The figure which shows an example of the search result information memorize | stored in a memory | storage means The figure which shows an example of the map screen displayed at the time of application starting The figure which shows an example of the map screen displayed after application starting The figure which shows an example of the image reproduction screen which is displayed after starting the application The figure which shows an example of the map screen displayed at the time of a motion detection scene reproduction | regeneration The figure which shows the method of searching the motion detection scene from the image of each camera Flow chart showing video storage process including event detection Flow chart showing start process of video search application Flowchart showing search process of motion detection scene Detailed flowchart of event pattern motion detection scene search processing Detailed flowchart of event search processing Flow chart showing video playback process Diagram showing the positional relationship between the camera icon and the playback cursor Flow chart showing video playback process using playback cursor The figure which shows an example of the map screen of Embodiment 2. The figure which shows an example of the map screen of Embodiment 3. The figure which shows an example of the map screen of Embodiment 4. The figure which shows an example of the map screen of Embodiment 5. FIG. 9 is a block diagram showing a configuration of a multi-camera video search device according to a sixth embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Camera 12 Multi-camera image search apparatus 13 Display 14 Mouse 21 Storage means 22 Search means 23 Reproduction means 24 Event detection means 25 Video input means 26 Map display means 27 Video output means 28 Condition input means 53 Arrow 54, C1-C3 Playback cursor

Claims (10)

A multi-camera video search device for searching video from a plurality of cameras,
Condition input means for designating the plurality of cameras and their order;
A moving object detection scene searching unit for searching a moving object detection scene in which a moving object is detected according to the plurality of designated cameras and the order;
A multi-camera video search apparatus characterized by comprising: And a map display means for displaying a map on which camera icons indicating the positions of the plurality of cameras are displayed.
The condition input means designates the plurality of cameras and their order by connecting the camera icons on the map with arrows.
The multi-camera video search device according to claim 1, wherein The map display means calculates the connection order of camera icons to be searched each time a plurality of camera icons are connected by arrows, and displays numerical values beside each camera icon.
The multi-camera video search device according to claim 2, wherein: The map display means includes
When the motion detection scene searched by the search means is played back, the camera icon that captured the motion detection scene being played back is highlighted.
The multi-camera video search device according to claim 2, wherein: The map display means includes
When the motion detection scene searched by the search means is played back, the imaging field of view range of the camera icon that shot the motion detection scene being played back is superimposed on the map,
The multi-camera video search device according to claim 2, wherein: The map display means includes
When the moving object detection scene retrieved by the retrieval unit is reproduced, the reproduction cursor indicating the reproduction time is moved and displayed on the arrow input by the condition input unit toward the camera icon in which the moving object appears newly. ,
The multi-camera video search device according to claim 2, wherein: The map display means includes
When the moving object detection scene searched by the search unit is played, the playback cursor indicating the playback time is moved toward the camera icon in which the moving object newly appears on the arrow input by the condition input unit. The moving object detection scene is displayed while being moved, and the moving cursor detection scene is reproduced on the reproduction cursor.
The multi-camera video search device according to claim 2, wherein: The map display means includes
The reproduction cursor indicating the reproduction time can be set at an arbitrary position on the arrow input by the condition input unit, and the reproduction time based on the position on the arrow input by the condition input unit at which the reproduction cursor is positioned Playing the searched motion detection scene of
The multi-camera video search device according to claim 2, wherein:   A multi-camera image search / playback method, wherein a plurality of cameras specified from a plurality of camera images and a moving object detection scene in which a moving object is detected according to an order are searched.   A multi-camera video search program for causing a computer to search a plurality of cameras specified from a plurality of camera videos and a moving object detection scene in which a moving object is detected according to an order.

Images