JP2020022083A - Detection apparatus and method thereof, and program - Google Patents

Abstract

To provide a detection apparatus for detecting a failure in real time from an image constituted of moving image or a still picture with a varying background image.SOLUTION: Shooting means shoots an image constituted of at least either one of a moving image and a still picture. Control means controls a shooting state of the shooting means to periodically vary. A detection apparatus includes: a shooting incidental information recording section for recording an image shot by the shooting means and shooting time; and a detection section for comparing images shot by the shooting means in the same shooting state at different time on the basis of shooting time to detect some failure.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a detection device, a method, and a program for detecting an abnormality from an image including at least one of a moving image and a still image.

  The conventional automatic monitoring by a camera continuously captures an image in a fixed direction and an angle of view, and detects an abnormality based on a temporal change of the image (see Non-Patent Document 1).

"How to Use Background Subtraction Methods", [online], [Search July 9, 2018], Internet <URL: https://docs.opencv.org/master/d1/dc5/tutorial_background_subtraction.html>

  However, in the method of the related art, it is necessary to fix a position where an image is captured by the camera, and there is a problem that a monitoring range per camera is significantly reduced as compared with a case where the camera is turned to look around. .

  For example, if the facilities and equipment are long and the monitoring range is wide, such as seawalls and harbors installed on coasts and rivers, or airport fences, a large number of cameras will be installed for fixed camera operation with a narrow monitoring range. Must be implemented, which hinders the introduction of automatic monitoring systems.

  However, in the turning camera operation, since the background image changes, it is difficult to detect an abnormality based on a temporal change of the video. Therefore, when monitoring a facility or the like having a wide monitoring range using a revolving camera, it is necessary for a human to detect an abnormality from the moving image. In practice, always checking videos is too expensive for the frequency of anomalies. It is common practice to check. For the reasons described above, it is difficult to realize a method of automatically detecting an abnormality in real time from a moving image captured using a revolving camera. Here, the abnormal condition means a state different from the normal state, for example, a state where some kind of moving object is entering a place where access is originally restricted or a place with little traffic, or a construction such as a dike. This is a state in which an object is destroyed by some factor (collision of a ship, aging, etc.), and is a state that can be detected by a temporal change of a moving image.

  An object of the present invention is to provide a detection device, a method, and a program for detecting an abnormality in real time from an image composed of a moving image or a still image in which a background image changes.

  According to one embodiment of the present invention, in order to solve the above-described problem, the image capturing unit captures an image including at least one of a moving image and a still image, and the control unit controls the image capturing so as to change periodically. The detecting device controls the photographing state of the means, and the detecting device performs photographing with the photographing means having the same photographing state based on the photographing time and the photographing incidental information recording unit that records the image photographed by the photographing means and the photographing time. And a detector for comparing images at different times and detecting an abnormality.

  In order to solve the above problem, according to another aspect of the present invention, the photographing unit is configured to photograph an image including at least one of a moving image and a still image, and the control unit is configured to change periodically. The detecting device controls the photographing state of the photographing means, and the detecting device includes: a photographing auxiliary information recording unit that records an image photographed by the photographing means and a photographing state of the photographing means at the time of photographing; And a detector for detecting an abnormality by comparing images at different times photographed by photographing means in the same photographing state based on the image data.

  ADVANTAGE OF THE INVENTION According to this invention, there exists an effect that abnormality can be detected in real time from the image which consists of a moving image or a still image whose background image changes.

FIG. 2 is a functional block diagram of the wide area automatic monitoring system according to the first embodiment. FIG. 6 is a diagram showing an example of a processing flow of the wide area automatic monitoring system according to the first embodiment. The figure which shows the example of a change of a photography state. FIG. 7 is a diagram for describing an example of detecting an abnormality in a detection unit. FIG. 7 is a diagram for describing an example of detecting an abnormality in a detection unit. FIG. 9 is a diagram for explaining processing contents of a detection unit of Modification 3 according to the first embodiment.

  Hereinafter, embodiments of the present invention will be described. In the drawings used in the following description, components having the same functions and steps for performing the same processing are denoted by the same reference numerals, and redundant description will be omitted.

<First embodiment>
FIG. 1 is a functional block diagram of the wide area automatic monitoring system according to the first embodiment, and FIG. 2 shows a processing flow thereof.

  The wide area automatic surveillance system includes a swivel type surveillance camera 10 with a pre-programmed automatic swivel control function and a detection device 20.

<Monitoring camera 10>
The turning type surveillance camera 10 with an automatic turning control function of a pre-programming type is mounted on a gantry, and moves the direction of the lens left and right (pan) and moves the direction of the lens up and down according to the contents programmed in advance. You can move (tilt), telephoto or wide-angle (zoom). In other words, the left and right angles of the lens, the up and down angles of the lens, and the angle of view of the camera can be controlled according to the contents programmed in advance. The left and right angles and the up and down angles of the lens and the angle of view of the camera are also referred to as a shooting state of the camera. For example, a control signal is sent to a stepping motor mounted on a gantry, actuated to control the left and right angles and up and down angles of the lens, a control signal is sent to the camera, and the angle of view is controlled by optical zoom or digital zoom I do.

  The program for controlling the camera is stored, for example, in a turning type monitoring camera 10 having an automatic turning control function of a pre-programming type, or in a storage unit (not shown) in the detecting device 20. A control unit (not shown) in the monitoring camera 10 or the detection device 20 retrieves the program from the storage unit, executes the program, and controls the shooting state of the monitoring camera 10 so as to change periodically according to the program.

The surveillance camera 10 may be controlled to photograph a predetermined position by prior programming. In order to repeat the same processing by the program, when a predetermined time has elapsed from the start of the program or the initial photographing state, control is performed so that a predetermined photographing state is obtained (photographing a predetermined position). In other words, a predetermined time corresponding to a predetermined shooting state is determined. Various methods can be considered as a method of controlling the shooting state of the camera based on the contents programmed in advance. For example,
(i) Determine the route and the moving speed of the lens and the changing speed of the angle of view in advance, and shoot according to it,
(ii) Determine one or more predetermined shooting conditions in advance, and shoot in each predetermined shooting condition,
Control. According to the method (i), the shooting state of the camera and its image can be specified from the elapsed time, and in the case (ii), a certain shooting state (for example, an initial state) is previously changed to a predetermined shooting state. By measuring the time it takes to achieve, it is possible to identify a predetermined shooting state and its image from the elapsed time. In addition, (A) the elapsed time may be reset in a certain shooting state (for example, an initial state or each predetermined shooting state), and an image may be acquired with the elapsed time from the shooting state, or (B) the elapsed time may be reset. The image may be obtained with the elapsed time from the start of shooting without starting (because the cycle does not change, for example, it is possible to specify a predetermined shooting state and the image from a remainder obtained by dividing the elapsed time by the cycle. it can).

  In the example of FIG. 3, it is assumed that the lens is moved as indicated by an arrow 90 and seven images t-1 to t-7 are acquired in one cycle. Note that t is an index indicating a cycle, and images obtained in the t-th cycle are represented as t-1 to t-7. For images t-1 to t-3, the direction of the lens is moved down and changed to a wide angle. For the images t-3 to t-7, the direction of the lens is moved right. After photographing the image t-7, the process returns to the image t-1 as indicated by an arrow 98, and shifts to the next cycle. It should be noted that the lens may be moved as indicated by an arrow 99 to acquire images t-8 to t-14 corresponding to the images t-7 to t-1, and a configuration in which 14 images are acquired in one cycle. Good.

  The monitoring camera 10 captures an image of the monitoring target (S10), and outputs the captured moving image. Note that not all of the images constituting the moving image need be output, but only a part of them may be output. For example, a moving image includes at least images t-1 to t-7.

<Detection device 20>
The detection device 20 includes a shooting incidental information recording unit 21 and a detection unit 23.

  The detection device 20 receives a moving image captured by the monitoring camera 10 as an input, detects an abnormality in the moving image, and outputs a detection result. In addition, the monitoring camera 10 and the detection device 20 may be connected by wire or may be connected by wireless. Also, the connection may be made directly or through another device. Further, the detection device 20 may be built in the monitoring camera 10.

  The detection device is, for example, a special device configured by reading a special program into a known or dedicated computer having a central processing unit (CPU), a main storage device (RAM: Random Access Memory), and the like. It is. The detection device executes each process under the control of the central processing unit, for example. The data input to the detection device and the data obtained in each process are stored in, for example, a main storage device, and the data stored in the main storage device is read out to a central processing unit as necessary, and is stored in another processing device. Used for processing. At least a part of each processing unit of the detection device may be configured by hardware such as an integrated circuit. Each storage unit included in the detection device can be configured by, for example, a main storage device such as a RAM (Random Access Memory) or a middleware such as a relational database or a key-value store. However, each storage unit does not necessarily need to be provided in the detection device, but is constituted by an auxiliary storage device including a semiconductor memory element such as a hard disk, an optical disk, or a flash memory, and is provided outside the detection device. May be provided.

  Hereinafter, each unit will be described.

<Shooting incidental information recording unit 21>
The shooting supplementary information recording unit 21 stores the moving image and the time (for example, a time stamp) at the time of shooting the moving image (S21). For example, at least the images t-1 to t-7 included in the moving image and the time (time stamp) at the time of capturing the image are stored.

  When the monitoring camera 10 outputs the time at the time of shooting a moving image together with the moving image, the shooting supplementary information recording unit 21 receives the time at the time of shooting the moving image together with the moving image as an input, and combines the time at the time of shooting the moving image with the moving image. Is stored in the photographing incidental information recording unit 21. In a configuration in which the monitoring camera 10 outputs only a moving image and does not output the time at the time of shooting a moving image, the shooting incidental information recording unit 21 sets the time at which the moving image was received as the time at which the moving image was shot, and Is stored in the photographing supplementary information recording unit 21. In this case, it is assumed that the detection device 20 includes a unit for obtaining a time (for example, an NTP (Network Time Protocol) client function), and the imaging incidental information recording unit 21 can obtain the time when the moving image is received.

<Detector 23>
The detection unit 23 extracts the combination of the moving image and the time at the time of shooting the moving image from the shooting incidental information recording unit 21, compares images captured by the monitoring camera 10 in the same shooting state at different times based on the shooting time, and determines whether there is an abnormality. It detects (S23) and outputs the detection result.

  As described above, when a predetermined time has elapsed from the start of the program control or the initial shooting state, the camera is controlled so as to be in the predetermined shooting state. Images at different times can be specified. For example, in the case of FIG. 3, if the initial photographing state is a photographing state in which the image t-1 can be acquired, the lens is moved from the initial photographing state as shown by an arrow 90, and the image is taken after s1 (for example, 10) seconds in the first cycle. 1-1, an image 1-2 is obtained after s2 (for example, 20) seconds, an image 1-3 is obtained after s3 (for example, 30) seconds, and an image 1-3 is obtained after s4 (for example, 40) seconds. 4, an image 1-5 is obtained after s5 (for example, 50) seconds, an image 1-6 is obtained after s6 (for example, 60) seconds, and an image 1-7 is obtained after s7 (for example, 70) seconds. Then, the lens is moved as indicated by an arrow 98, and similarly, images 2-1 to 2-7 are acquired s1 to s7 seconds after the time of the initial shooting state in the second cycle. In this case, the image ti and the image t'-i are images captured at different times by the monitoring camera 10 in the same capturing state, and two images can be specified based on the capturing time. Here, i is an index indicating an image, i = 1, 2,..., 7, and t ′ is an index indicating a cycle, and is t′t ′. Note that this is an example of a method of specifying images at different times captured by the monitoring camera 10 in the same capturing state based on the capturing time, and various specifying methods may be considered based on the control method. For example, when the lens is moved from the lens position of the image t-7 as indicated by an arrow 99 to acquire images t-8 to t-14 corresponding to the images t-7 to t-1, the images ti and image t′-i is an image taken at a different time, which was taken by the monitoring camera 10 in the same shooting state, and further, the image ti and the image t- (14-i + 1) were taken by the monitoring camera 10 in the same shooting state. , Images at different times.

  In the conventional turning-type operation, since the background image changes, it is difficult to detect an abnormality based on a temporal change of the video. Therefore, in the present embodiment, attention is paid to a periodically changing background image, and a change in the image is detected at the timing when the same background image is obtained in the periodically changing background image, thereby detecting an abnormality.

  When comparing images captured at different times by the monitoring camera 10 in the same shooting state, the image (the latest image) acquired in real time is compared with the comparison source image. Can be considered. Hereinafter, two examples of the method of acquiring the comparison source image will be described.

(Acquisition method 1)
In advance, the monitoring camera 10 is automatically patrol-photographed by a program, and the moving image and the time at which the moving image was captured are acquired as comparison source images. For example, when monitoring a building or the like, by acquiring a comparison source image immediately after building, a change due to aging can be detected as abnormal. If the background image changes depending on the season or time (for example, morning, noon, or night), the comparison source image is acquired for each season or time, and the comparison is performed according to the acquired season or time for the image acquired in real time. The original image may be changed.

  All frames (all images) of a moving image may be acquired as comparison source images, or some frames (partial images) may be extracted as comparison source images as shown in FIG. It should be noted that the extraction ratio may be appropriately set according to the required accuracy, and may be extracted at a ratio at which the entire monitoring target can be checked at least. In the present embodiment, a part of the image is extracted as shown in FIG.

(Acquisition method 2)
An image acquired in the past in real time is set as a comparison source image. For example, the image (t-1) -i one cycle before is set as the comparison source image. Also, the average or weighted average of u or more images (t-1) -i, (t-2) -i,..., (Tu) -i acquired in real time in the past (for example, (Increase the weight) may be used as the comparison source image. Further, the comparison source image may be updated each time a new image is acquired, or may be updated every predetermined time interval (1 minute, 10 minutes, 1 hour, 1 day). With such a configuration, the trouble of separately acquiring the comparison source image as in the acquisition method 1 can be omitted. In addition, the trouble of acquiring the comparison source image for each season or each time can be omitted. Note that in this acquisition method, as described in the acquisition method 1, all frames (all images) of the moving image may be acquired as comparison source images, or some frames (partial images) as shown in FIG. May be extracted as the comparison source image. At this time, when extracting a part of the images, the time at which the comparison source image to be extracted was photographed may be specified in advance.

  Regardless of which method the comparison source image is obtained, the abnormality can be detected. However, when the building is gradually deformed due to a change due to aging or the like, it is easier to detect a state abnormality by using the comparison source image acquired by the acquisition method 1. In the case of the acquisition method 2, since the deformed state is used as the comparison source image, the gradually occurring deformation may not be detected in some cases.

  Note that the comparison source image is stored in a storage unit (not shown) or the shooting incidental information recording unit 21.

(Detection processing)
The detection unit 23 uses the comparison original image obtained by the above-described method and the combination of the moving image taken out from the shooting incidental information recording unit 21 and the time at the time of shooting the moving image to be shot by the monitoring camera 10 in the same shooting state. Compare images at different times. For example, a difference between two images is obtained, and if the difference exceeds a predetermined threshold, the image is detected as abnormal and a detection result is output. For example, when an abnormality occurs as shown in FIG. 4, the difference between the comparison source image 0-4 and the image t-4 acquired in real time becomes large and is detected as abnormal (see FIG. 5). In addition, as a method of “comparing images”, a method of comparing feature amounts obtained from images, and the like can be considered in addition to a method of directly comparing images. In the case of the method of comparing the feature amounts, the feature amounts are extracted from the comparison source image, stored, the feature amounts are extracted from the image acquired in real time, and the feature amounts are compared to compare the images. I do.

(Output of detection result)
As a method of outputting the detection result, any method may be used. For example, a moving image captured by the surveillance camera 10 may be displayed on a display unit such as a display, and a message or image notifying that an abnormality has been detected in the moving image may be superimposed and displayed. It may be displayed (enclosed in a bold line or the like of a color not used in the moving image). Further, a warning sound or the like notifying that an abnormality has been detected may be emitted from a speaker, a rotating light or the like notifying that an abnormality has been detected may be operated, or a combination thereof may be used.

<Effect>
With the above configuration, an abnormality can be detected in real time from an image included in a moving image in which the background image changes. The user of this monitoring system can know the abnormality in real time by always using a display, a warning sound, a rotating light, and the like without constantly checking a moving image.

<Modification 1>
The following description focuses on the differences from the first embodiment.

  In the present embodiment, the revolving surveillance camera 10 with the pre-programming type automatic revolving control function has been described as being capable of controlling the shooting state without temporal fluctuation according to the contents programmed in advance. . However, at the time of actual shooting, mechanical problems such as external factors such as wind and mechanical accuracy and aging deterioration of the revolving surveillance camera 10 having an automatic revolving control function (hereinafter also referred to as external factors, etc.) ), Temporal fluctuation may occur in the automatic turning operation. For example, the normal cycle (the cycle at the time of acquiring the comparison source image) may be 90 seconds or 110 seconds due to external factors or the like, despite being 100 seconds. In this case, the comparison source image and the image obtained in real time cannot be appropriately compared. For example, if the frame rate of the moving image is 30 fps and the cycle when the comparison source image is obtained is 100 seconds, 3000 images are obtained in one cycle. When the period of the images acquired in real time becomes 90 seconds and 110 seconds due to an external factor or the like, 2700 images and 3300 images are acquired, respectively. At this time, there is a possibility that different positions may be photographed between the image acquired in real time y seconds after the start of a certain cycle and the comparison source image y seconds after the start of the cycle. Can not.

The detection unit 23 of the present modification corrects the time y at the time of capturing the image acquired in real time from the period a when acquiring the image in real time and the period b when acquiring the comparison source image. For example, if the time after correction is y ',
y '= y × (b / a)
To obtain the corrected time y ′. It should be noted that the turning type monitoring camera 10 or the detecting device 20 with the automatic turning control function of the pre-programming type takes a time from a certain shooting state (for example, an initial shooting state) to another shooting state to return to a certain shooting state again ( Period) can be obtained.

<Effect>
With such a configuration, the same effect as in the first embodiment can be obtained. Further, even when a temporal fluctuation occurs in the automatic turning operation, an abnormality can be detected.

<Modification 2>
The following description focuses on the differences from the first embodiment.

  In the above-described first modification, the time y at the time of capturing the image acquired in real time is corrected. However, since the speed of the automatic turning operation is not always constant, the capturing state at the time after the correction is changed by the comparison source. It does not always match the shooting state of the image. Therefore, in this modification, the similarity between images is used.

  The detection unit 23 calculates the similarity between each comparison-source image and the image acquired in real time, acquires the one with the highest similarity to each comparison-source image, and compares the comparison-source image with the acquired image. I do.

  The detecting unit 23 extracts a combination of the moving image and the time at which the moving image was shot from the shooting incidental information recording unit 21. For example, if the frame rate of a moving image is 30 fps and the cycle is 100 seconds, 3,000 images are acquired in one cycle. When the period becomes 90 seconds or 110 seconds due to an external factor or the like, 2700 images and 3300 images are acquired, respectively.

  The detection unit 23 calculates the similarity between each comparison source image (all seven in the case of FIG. 3) and all the images acquired in real time (for example, 3000, 2700, or 3300), and calculates each comparison source image. , The image having the highest similarity to the acquired image may be acquired, and the comparison source image may be compared with the acquired image.

  If a range of temporal fluctuation can be assumed for each comparison source image, the similarity may be calculated only within the range, and the image having the highest similarity may be acquired. For example, the comparison source image 0-1 is obtained by moving the lens from the initial shooting state and after s1 (for example, 10) seconds, and moves to the shooting state of the comparison source image 0-1 from the initial shooting state. If the temporal fluctuation that can occur up to 1 second is less than 1 second, move the lens from the initial shooting state, and obtain an image acquired in real time within 1 second before and after s1 (for example, 10) seconds (in the initial shooting state Only the similarity between the 270th to 330th images counted from the corresponding image) and the comparison source image 0-1 is calculated, and the image with the highest similarity is obtained and compared with the comparison source image 0-1. Good.

  Further, since the comparison source images are arranged in chronological order, when an image ti to be compared with a certain comparison source image 0-i is obtained, the similarity is calculated with the next comparison source image 0- (i + 1). The image to be performed may be limited to an image temporally later than the image ti.

<Effect>
With such a configuration, the same effect as in the first embodiment can be obtained. Further, even when a temporal fluctuation occurs in the automatic turning operation, an abnormality can be detected. Furthermore, even when the speed of the automatic turning operation is not constant, it can be handled. However, depending on the monitoring target, as in the case of images t-5 and t-6 in FIG. 5, the images may be similar in spite of different image positions, and in such a case, the modification 1 It is better to choose a method.

<Modification 3>
The following description focuses on the differences from the first embodiment.

  When a temporal fluctuation occurs in the automatic turning operation, the time corresponding to the shooting time of the comparison source image is included in the combination of the moving image and the time at the time of shooting the moving image from the shooting incidental information recording unit 21 extracted by the detection unit 23. May not exist in real time. In such a case, the detection unit 23 (A) “the time elapsed from the time in the initial shooting state when acquiring the comparison source image to the acquisition time when acquiring the comparison source image” and “the image in real time If the difference from "the time elapsed from the time in the initial photographing state at the time of acquisition to the acquisition time at which the image is acquired" is smaller than a predetermined threshold or equal to or smaller than a predetermined threshold, the two images are compared. On the other hand, the detection unit 23 includes (B) “the time elapsed from the time in the initial shooting state when acquiring the comparison source image to the acquisition time when acquiring the comparison source image” and “the initial time when acquiring the image in real time. If the difference from the time in the shooting state to the time elapsed from the time of acquisition of the image to the acquisition time is equal to or greater than the predetermined threshold or greater than the predetermined threshold, The difference between the time elapsed before the acquisition time to acquire the comparison source image and the time elapsed from the time in the initial shooting state when acquiring the image in real time to the acquisition time to acquire the image is the two smallest. The images are combined, and the combined image and the comparison source image are compared (see FIG. 6). For example, two images are superimposed and combined so that the difference is minimized, and an image in which an unnecessary portion is deleted so as to be compared with the comparison source image is defined as the above-described combined image. At this time, the virtual shooting state corresponding to the synthesized image and the shooting state corresponding to the comparison source image are the same shooting state.

<Modification 4>
In the present modification, a control unit (not shown) in the monitoring camera 10 or the detection device 20 determines that the speed of change at a predetermined time corresponding to one or more predetermined shooting states is one or more predetermined shooting conditions. Control is performed so as to be slower than the speed of change at a predetermined time corresponding to a shooting state other than the state. For example, the moving speed of the lens at the time corresponding to the shooting state at which the images t-1 to t-7 are obtained is made slower than the moving speed at other times, and the quality of the images t-1 to t-7 to be obtained. Can be increased. Furthermore, in the programming performed in advance, the imaging state may be fixed when acquiring the images t-1 to t-7.

  At this time, control may be performed so that temporal fluctuations are absorbed by the fixed time. For example, when changing the shooting state from the shooting state in which the image t-1 in FIG. 3 can be obtained to the shooting state in which the image 1-2 can be obtained, it takes 10 seconds if there is no external factor or the like. If there is a delay of 2 seconds at the maximum when there is, etc., the shooting state is changed from the shooting state in which the image t-1 can be obtained to the shooting state in which the image 1-2 can be obtained, and the image t-1 Is controlled so as to acquire the image 1-2 after 12 seconds from the photographing state in which can be acquired. With such a configuration, even if there is an external factor or the like, the image 1-2 can be acquired after the photographing state in which the image 1-2 can be acquired.

  In the present embodiment, the moving speed of the image needs to be suppressed to a human-recognizable speed so that the user can check the monitoring location corresponding to the images t-1 to t-7 on one display. In the embodiment, seven displays may be prepared, and the images t-1 to t-7 may be displayed on each display. In this case, the imaging state is stopped in the imaging state for acquiring the images t-1 to t-7, and in the other imaging states, the imaging state is changed at a high speed regardless of whether or not a human can recognize it. For example, the period is set to one second, and images t-1 to t-7 acquired every second are displayed on seven displays. In the conventional fixed camera operation, seven cameras need to be installed, but in this configuration, one camera may be installed. In this case, the imaging may be stopped in an imaging state other than the imaging state for acquiring the images t-1 to t-7 (that is, in a state in which the direction of the lens and the angle of view are changed).

<Other modifications>
In the present embodiment and its modified examples, the description has been made assuming that a moving image is photographed. However, a similar effect can be obtained by photographing a still image. In short, images at different times, which are captured by cameras having the same capturing state, may be used.

  In the present embodiment, some images included in a moving image (an expression style using a beta motion in which a still image appears to be moving as an illusion of human vision when a continuously changing still image is continuously switched at high speed) Although the detection process is performed based on (frame) ti, the detection process may be performed based on all images (frames) ti included in the moving image.

  In the present embodiment, the shooting state of the camera is set by matching the left and right angles of the lens, the vertical angle of the lens, and the angle of view of the camera, but at least the left and right angles of the lens, the vertical angle of the lens, and the angle of view of the camera. Either one may be changed. Therefore, the shooting state may be information indicating at least one of the states of the left and right angles of the lens, the up and down angles of the lens, and the angle of view of the camera (a state that changes).

<Second embodiment>
The following description focuses on the differences from the first embodiment.

  Even by the methods described in Modifications 1 to 3 of the first embodiment, errors due to temporal fluctuations cannot always be eliminated. Therefore, in the present embodiment, an image acquired in real time is specified using the shooting state of the camera at the time of shooting.

  The swivel-type surveillance camera 10 with an automatic swivel control function of a pre-programming type, based on a control signal indicating a predetermined shooting state, sets the left and right angles, the up and down angles of the lens, and the angle of view of the camera so that the shooting state is achieved. Control.

  In this embodiment, the surveillance camera 10 completes (i) changing the shooting state of the monitoring camera 10 itself to a predetermined shooting state indicated by the control signal at a predetermined cycle or every time the shooting state is changed. Then, when the signal indicating completion (not the shooting state of the monitoring camera 10 itself) or (iii) the change to the predetermined shooting state indicated by the control signal is completed, the shooting of the monitoring camera 10 itself at the time of completion is completed. It shall output the status. In the above cases (i) and (iii), the monitoring state of the monitoring camera 10 itself output by the monitoring camera 10 is used, and in the above (ii), the control signal is used to The photographing state of No. 10 is specified (essentially, processing is performed assuming that the monitoring camera 10 is in the photographing state indicated by the control signal). The control signal may be output from the detection device 20 or may be output from the monitoring camera 10.

<Shooting incidental information recording unit 21>
The shooting supplementary information recording unit 21 stores the moving image and the shooting state of the camera at the time of shooting the moving image (S21). For example, at least the images t-1 to t-7 of FIG. 3 included in the moving image and the photographing state of the camera at the time of photographing the images are stored. Note that, as in the first embodiment, the shooting time of the camera at the time of shooting a moving image may be included.

  When the surveillance camera 10 outputs the shooting state of the camera at the time of shooting a moving image together with the moving image (in the cases (i) and (iii) described above), the shooting incidental information recording unit 21 The shooting state of the camera is received as an input, and a combination of the moving image and the shooting state of the camera at the time of shooting the moving image is stored in the shooting incidental information recording unit 21. In the case where the monitoring camera 10 outputs only a signal indicating that the change to the predetermined shooting state indicated by the moving image and the control signal is completed, and does not output the shooting state at the time of shooting the moving image (in the case of the above (ii)) First, the shooting incidental information recording unit 21 sets the predetermined shooting state indicated by the control signal as the shooting state at the time of moving image shooting, and stores a combination of the moving image and the time at the time of moving image shooting in the shooting incidental information recording unit 21.

<Detector 23>
The detection unit 23 extracts the combination of the moving image and the shooting state from the shooting incidental information recording unit 21, compares the images captured by the monitoring camera 10 in the same shooting state at different times based on the shooting state, and detects an abnormality (S23). ), And output the detection result.

  In the present embodiment, it is possible to acquire the photographing state itself and specify images captured at different times by the monitoring camera 10 having the same photographing state. For example, in the case of FIG. 3, assuming that the initial photographing state is a photographing state in which the image t-1 can be acquired, the lens is moved from the initial photographing state as indicated by an arrow 90, and regardless of the photographing time, each photographing state at the t-th week. , Images t-1 to t-7 are obtained. In this case, the image ti and the image t'-i are images captured at different times by the monitoring camera 10 in the same capturing state, and two images can be specified based on the capturing time. Note that this is an example of a method of specifying images at different times captured by the monitoring camera 10 in the same capturing state based on the capturing time, and various specifying methods may be considered based on the control method. For example, when the lens is moved from the lens position of the image t-7 as indicated by an arrow 99 to acquire images t-8 to t-14 corresponding to the images t-7 to t-1, the images ti and image t′-i is an image taken at a different time, which was taken by the monitoring camera 10 in the same shooting state, and further, the image ti and the image t- (14-i + 1) were taken by the monitoring camera 10 in the same shooting state. , Images at different times.

  With such a configuration, similarly to the first embodiment, it is possible to detect a change in an image at a timing at which the same background image is obtained in a periodically changing background image, and to detect an abnormality.

  When comparing images captured at different times by the monitoring camera 10 in the same shooting state, the image obtained in real time (the latest image) is compared with the comparison source image. Various things are conceivable as in one embodiment. The comparison source image can be acquired by the same acquisition method as in the first embodiment by using the shooting state at the time of shooting a moving image instead of the time at the time of shooting a moving image. The detection process and the output of the detection result are the same as in the first embodiment.

<Effect>
With the above configuration, the same effects as in the first embodiment can be obtained. Furthermore, in the present embodiment, since the image acquired in real time is specified using the photographing state of the camera at the time of photographing, there is no need to consider temporal fluctuation. Note that the present embodiment may be combined with Modification 4 of the first embodiment and other modifications.

<Modification>
The following description focuses on the differences from the second embodiment.

  In the present embodiment, a moving image is shot by using a pre-programming type turning surveillance camera 10 having an automatic turning control function. In this modification, a pre-programming type monitoring camera 10 having an auto-pilot function is used. . The surveillance camera 10 is mounted on an unmanned aerial vehicle, and navigates and changes the position according to pre-programmed contents, moves the lens direction left and right, moves the lens direction up and down, and performs telephoto or wide-angle. can do. In other words, it is possible to control the position of the monitoring camera 10 itself, the left and right angles of the lens, the up and down angles, and the angle of view of the camera according to the contents programmed in advance. In the present embodiment, the position of the monitoring camera 10 itself, the left and right angles and the up and down angles of the lens, and the angle of view of the camera are also referred to as a camera shooting state. For example, by transmitting a control signal to an unmanned aerial vehicle, controlling the position and attitude of the unmanned aerial vehicle, controlling the position of the surveillance camera, the left and right angles of the lens, and the up and down angles, transmitting a control signal to the camera, The angle of view is controlled by zoom or digital zoom. The surveillance camera 10 with the autopilot function can acquire its own position using a GPS (Global Positioning System) or the like, and can use its own gyro sensor or geomagnetic sensor to detect its own posture (the left and right angles of the lens). , Upper and lower angles), and its own angle of view can be obtained from the angle of view control mechanism.

  The program for controlling the camera is stored, for example, in the monitoring camera 10 with an auto-pilot function of a pre-programming type or in a storage unit (not shown) in the detection device 20. A control unit (not shown) in the monitoring camera 10 or the detection device 20 retrieves the program from the storage unit, executes the program, and controls the shooting state of the monitoring camera 10 so as to change periodically according to the program.

  The surveillance camera 10 may be controlled to photograph a predetermined position by prior programming.

  In the case of an unmanned aerial vehicle, temporal fluctuations are more likely to occur due to external factors and the like than in a surveillance camera mounted on a gantry and turning. Therefore, the detection unit 23 detects an abnormality by comparing images captured at different times by the monitoring camera 10 in the same shooting state based on the shooting state, not the shooting time. However, if control can be performed so as to be in a predetermined shooting state at a predetermined time, the detection unit 23 uses the monitoring camera 10 in the same shooting state based on the shooting time, as in the first embodiment. Abnormalities may be detected by comparing captured images at different times.

<Effect>
With such a configuration, the same effect as in the second embodiment can be obtained. Further, the use of the monitoring camera 10 with an auto-pilot function of a pre-programming type enables flexible change of the route.

<Other modifications>
The present invention is not limited to the above embodiments and modifications. For example, the above-described various processes may be executed not only in chronological order as described, but also in parallel or individually according to the processing capability of the device that executes the processes or as necessary. In addition, changes can be made as appropriate without departing from the spirit of the present invention.

<Program and recording medium>
Further, various processing functions in each device described in the above embodiment and the modified examples may be realized by a computer. In this case, the processing content of the function that each device should have is described by a program. By executing this program on a computer, various processing functions of the above-described devices are realized on the computer.

  A program describing this processing content can be recorded on a computer-readable recording medium. As a computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  The distribution of the program is performed by, for example, selling, transferring, lending, or the like, a portable recording medium such as a DVD or a CD-ROM on which the program is recorded. Further, the program may be stored in a storage device of a server computer, and the program may be distributed by transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its own storage unit. Then, when executing the processing, the computer reads the program stored in its own storage unit and executes the processing according to the read program. Further, as another embodiment of the program, the computer may directly read the program from the portable recording medium and execute processing according to the program. Further, each time a program is transferred from the server computer to the computer, processing according to the received program may be sequentially performed. A configuration in which the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes a processing function only by executing an instruction and acquiring a result without transferring a program from the server computer to the computer. It may be. It should be noted that the program includes information to be used for processing by the computer and which is similar to the program (such as data that is not a direct command to the computer but has properties that define the processing of the computer).

  Further, each device is configured by executing a predetermined program on a computer, but at least a part of the processing contents may be realized by hardware.

Claims (8)

The photographing means shall photograph an image composed of at least one of a moving image and a still image,
The control means controls the photographing state of the photographing means so as to periodically change,
An image captured by the image capturing means, and a capturing auxiliary information recording unit that records the capturing time,
Based on the photographing time, photographed by the photographing means in the same photographing state, including a detection unit that compares the images at different times and detects an abnormality.
Detection device. The photographing means shall photograph an image composed of at least one of a moving image and a still image,
The control means controls the photographing state of the photographing means so as to periodically change,
An image photographed by the photographing unit, and a photographing incidental information recording unit that records a photographing state of the photographing unit at the time of photographing;
A detection unit that compares the images at different times and detects an abnormality, based on the imaging state of the imaging unit at the time of the imaging, and captured by the imaging unit in the same imaging state.
Detection device. The detection device according to claim 1 or 2, wherein
The control means may be a predetermined time corresponding to one or more predetermined photographing states, or the speed of change of the photographing state of the photographing means in one or more predetermined photographing states may be the one or more predetermined photographing states. At a predetermined time corresponding to a shooting state other than the shooting state, or controlled to be slower than the speed of change of the shooting state of the shooting unit in the one or more shooting states other than the predetermined shooting state,
The detection unit detects an abnormality by comparing the images at different times, photographed by the photographing unit in the one or more predetermined photographing states,
Detection device. The detection device according to any one of claims 1 to 3, wherein
The control means controls a shooting state of the shooting means by controlling a gantry on which the shooting means is mounted,
Detection device. The detection device according to any one of claims 1 to 3, wherein
The control means controls a shooting state of the shooting means by controlling an unmanned aerial vehicle equipped with the shooting means,
Detection device. The photographing means shall photograph an image composed of at least one of a moving image and a still image,
The control means controls the photographing state of the photographing means so as to periodically change,
A photographing incidental information recording step of recording the image photographed by the photographing means and the photographing time in a photographing incidental information recording unit,
A detection step of comparing the images at different times and detecting an abnormality, the images being captured by the imaging unit in the same imaging state based on the imaging time.
Detection method. The photographing means shall photograph an image composed of at least one of a moving image and a still image,
The control means controls the photographing state of the photographing means so as to periodically change,
A photographing supplementary information recording step of recording an image photographed by the photographing means and a photographing state of the photographing means at the time of photographing in a photographing supplementary information recording unit;
A detecting step of comparing the images at different times and detecting an abnormality, which is performed by the photographing unit in the same photographing state, based on a photographing state of the photographing unit at the time of the photographing.
Detection method.   A program for causing a computer to function as the detection device according to any one of claims 1 to 5.

Images