JP4803376B2 - Camera tampering detection method - Google Patents

Description

  The present invention relates to a camera tampering detection method, and more particularly to a method for detecting camera tampering by image processing.

  In a security device using a surveillance camera, it is difficult for a surveillance staff to constantly check the surveillance camera image monitor. For this reason, many monitoring apparatus systems that record and store only video in which an abnormality has occurred are considered. Examples of abnormal phenomena include intruders, camera obstructions, and lighting changes. There are various types of camera interference. When the signal line of the camera is cut, it is detected by the absence of the signal from the image signal line. When the power line is disconnected, a warning is issued based on the disconnection. Below are some examples of prior art related to this.

  The “photographing disturbance detection method for the surveillance camera” disclosed in Patent Document 1 is a method for automatically detecting and coping with the imaging disturbance for the surveillance camera. As shown in FIG. 3 (a), a first monitoring device comprising a CPU, a ROM, a RAM, an arithmetic processor, a frame memory, and a communication control unit compresses a still image of a monitoring object photographed by a monitoring camera. And transmitted as image data. A second monitoring device consisting of a CPU, ROM, RAM, arithmetic processor, frame memory, display device, communication control unit, and alarm device. Receives and decompresses image data from the first monitoring device via a communication line. Processing is performed to restore and display the still image of the monitored object for monitoring. When it is detected that the data amount of the image data obtained by compressing the image captured by the monitoring camera by the first monitoring device is smaller than a preset threshold value of the data amount, the image is captured by the monitoring camera. The occurrence of the abnormality is displayed / alarmed by the second monitoring device.

  The “surveillance camera system” disclosed in Patent Document 2 detects and outputs an alarm when the visual field of the surveillance camera is obstructed. As shown in FIG. 3 (b), the disturbance detection means detects whether or not the field of view of the imaging means is obstructed. Based on the detection result of the interference detection means, the alarm means outputs an alarm.

  The “image monitoring apparatus” disclosed in Patent Document 3 is an image monitoring apparatus that can self-diagnose whether an abnormality has occurred in an imaging unit or a motion detection unit or a mischief has occurred during monitoring of a monitoring area. As shown in FIG.3 (c), the inside of an elevator car is imaged with a CCD camera. The image data obtained by the CCD camera is compressed by the image compression unit. The compressed image data is stored in the image storage unit. A state change in the car is detected by the motion detector. Two pieces of compressed image data at different time points are fetched from the image compression unit or the image storage unit. Diagnosis is made by comparing the amount of image data with a diagnostic means. The motion detection unit compares the amount of compressed image data before and after the state change in the car is detected, and the imaging abnormality determination unit detects an abnormality of the CCD camera. If a change in the compressed image data amount is detected when no change in the state of the car is detected, the detection abnormality determination unit determines that the motion detection unit is abnormal.

  The “surveillance camera system” disclosed in Patent Document 4 detects an obstacle in the visual field of the surveillance camera and outputs an alarm. The disturbance detection means detects whether or not the field of view of the imaging means is obstructed. Based on the detection result of the interference detection means, the alarm means outputs an alarm. The interference detection means includes an illuminance sensor that detects the illuminance of the monitoring area. When the luminance value of the image picked up by the image pickup means is equal to or lower than a predetermined value and when the illuminance detected by the illuminance sensor is equal to or higher than the predetermined value, it is detected as a disturbing action.

  The “surveillance camera with disturbance detection function” disclosed in Patent Document 5 accurately detects a deterioration state of the resolution of a screen from a video signal without being affected by a change in brightness of a monitoring area. is there. As shown in FIG. 3D, a focus quality detection unit is provided that detects a change in the high-frequency contour signal due to a decrease in signal components of 2 MHz to 5 MHz in the video signal. It is possible to accurately detect a state where a translucent light-shielding object is attached to the lens surface of the surveillance camera.

  The “camera monitoring device” disclosed in Patent Document 6 is a camera monitoring device based on a background difference method that realizes a distinction between a true abnormality occurrence and an environmental change by a relatively simple process. As shown in FIG. 3E, the analog video signal from the camera is converted into a digital video signal by the A / D converter. It is stored in the background image memory unit and the input image memory unit of the image memory. Based on the data stored in the image memory, the image processing unit extracts a region where the difference between the input image and the background image exceeds a predetermined threshold. The time change of the area is accumulated in the system memory. Based on the data in each of the image processing unit and the system memory, the determination unit determines whether there is an abnormality. This determination unit determines whether the environmental change such as lighting, intruder, obstruction to the camera, video input interruption, etc. from the size of the area exceeding the predetermined threshold, its time change, etc. An alarm is issued when an abnormality occurs.

  The “surveillance camera tampering detection device” disclosed in Patent Document 7 stably detects tampering even in various scenes with different numbers of edges, even for a monitoring target in which the brightness changes only in part. Is. As shown in FIG. 3F, the image input unit captures the input image. An intruder detection unit detects an intruder from the input image. The reference image capturing unit automatically captures a reference image that is a criterion for determining whether or not the input image is disturbing. A reference image update timing generation unit provides an instruction signal for acquiring a reference image to the reference image acquisition unit. The image comparison unit compares the image features between each of the captured multiple reference images and the input image to determine the degree of inconsistency between the images, and outputs the lowest degree of inconsistency as the inconsistency of the input image. To do. Based on the magnitude and temporal change of the degree of inconsistency of the input image and the output signal of the intruder detection unit, it is determined by the interference determination unit whether the interference act is being performed.

JP 09-074553 JP 2000-222646 A JP 2000-293766 JP 2001-218189 A Japanese Patent Laid-Open No. 2001-238204 JP 2001-333417 A JP 2005-252479 A

  However, the conventional camera tampering detection method has the following problems. In the method of determining disconnection based on the presence / absence of a video signal, it is not possible to recognize a camera obstruction other than disconnection, such as an act of blocking the camera lens or an action of changing the direction. Further, since there are various abnormal phenomena, it is necessary to reproduce all the recorded images recorded as the abnormal phenomenon and actually confirm which phenomenon has occurred. In other words, it is necessary to confirm whether it is an intruder, camera obstruction, lighting change, etc., but in the conventional method, it is necessary to confirm in a short time what kind of camera obstruction occurred. Is not easy.

  An object of the present invention is to solve the above-described conventional problems so that it is possible to easily distinguish and confirm what kind of camera obstruction acts have occurred in a surveillance camera device or the like.

In order to solve the above problems, the present invention, the camera tampering detection method, a memory for holding the reference image, the luminance difference calculation unit for calculating a luminance difference by comparing the reference image and the image of the monitoring camera for each pixel When the change pixel extracting unit for counting the luminance difference by extracting the pixels of the above first threshold value as the change pixel, and the abrupt change detection unit for detecting a sudden change in the image, abrupt number and the image of the changed pixel A camera disturbance determination unit that determines the presence or absence of camera disturbance based on the presence or absence of a change, and a warning notification unit that issues a warning when camera disturbance is detected, and the camera disturbance determination unit includes: Means for determining that there is a possibility of camera interference when the number is greater than or equal to the second threshold, monochromatic interference detecting means for detecting monochromatic interference that changes the image to one color, and mixed change interference that changes the scene of the entire image Detect Mixing change disturbance detecting means, wherein the mixed change disturbance detecting means counts the number of bright change pixels that are brighter than the luminance obtained by adding a first threshold to the luminance of the pixels of the reference image; and Means for counting the number of dark change pixels less than or equal to the brightness obtained by subtracting the first threshold from the brightness of the pixels, and determining the presence or absence of mixing change interference based on the ratio of the number of bright change pixels to the number of dark change pixels And a means for carrying out.

  With the configuration described above, camera disturbance in a monitoring camera device or the like can be easily determined only by image processing. In other words, it is easy to check only the abnormal phenomenon that you want to check by judging whether the abnormal phenomenon on the image is due to foreign object intrusion, camera obstruction, or illumination change, using the image processing algorithm. become able to. In particular, it is possible to judge camera interference without misrecognizing it as a foreign object intrusion or a change in lighting, so that in the situation where camera interference is diversified, it is possible to correctly identify and handle each camera interference act appropriately. It becomes possible.

  Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to FIGS. 1 and 2. FIG.

  The embodiment of the present invention compares the image of the surveillance camera and the reference image to obtain a luminance difference for each pixel. If the luminance difference value is equal to or greater than the first threshold, the pixel is changed, and the changed pixel is equal to or greater than the second threshold. This is a camera tampering detection method that determines the type of camera tampering and issues a warning if a sudden change occurs.

  FIG. 1 is a functional block diagram of the camera tampering detection method in the embodiment of the present invention and a conceptual diagram showing a state of image change. FIG. 2 is a flowchart showing the processing procedure of the camera tampering detection method. In FIG. 1, the surveillance camera 1 is a television camera. In general, a two-dimensional optical sensor such as a digital still camera may be used. The image processing unit 2 is means for processing captured images to detect camera interference. The image acquisition unit 3 is means for acquiring an image from a camera. The luminance difference calculation unit 4 is a means for comparing the acquired image with a reference image to obtain a luminance difference for each pixel. The change pixel extraction unit 5 is a means for extracting change pixels having a luminance difference equal to or greater than a threshold value. The sudden change detection unit 6 is a means for determining that there is a sudden change in an image when the number of pixels having a luminance change equal to or greater than a threshold value between frames is equal to or greater than a predetermined value. The camera disturbance determination unit 7 is a means for statistically processing the luminance difference data to determine the type of camera disturbance. The reference image storage unit 8 is a memory that holds an image serving as a reference for comparison. The warning notification unit 9 is a means for notifying that camera interference has been detected.

  The function and operation of the camera tampering detection method in the embodiment of the present invention configured as described above will be described. First, an overview of the functions of the camera tampering detection method will be described. Obstructive actions for surveillance cameras include covering the camera lens with hands, painting with spray, applying adhesive such as gummed tape, changing the direction of the camera, and cutting the camera cord. There are roughly two types of camera obstructions on the image. The first monochromatic change is the image change seen during obstructions such as covering the camera lens with your hand, painting with spray, sticking adhesive such as gummed tape, cutting the camera cord, In this aspect, the image changes to a single color. The second mixed change is an image change that is observed during a disturbing action such as changing the direction of the camera, and is a mode in which a brighter portion and a darker portion become mixed changes compared to the reference image. . In the case of disturbances that change the screen to striped patterns or sandstorms, it is a mixed change as well.

  When such a disturbance is applied to the installed surveillance camera, it detects that a camera disturbance act has been performed by image processing and issues a warning. When camera obstruction is performed, it is compared with the reference image captured in advance to determine how it has changed from the reference image, so that a change in lighting or sudden sunshine change is not mistaken for camera obstruction. , Issue a warning by correctly judging only the camera obstruction. In a monitoring system using image processing such as intruder detection, camera interference can be determined only by image processing without using another device or means for detecting camera interference. In addition, by recording the video of the surveillance camera together with information such as the camera disturbance type, it is possible to easily detect only the occurrence of a specific abnormal condition by searching for the distinction between the type of camera disturbance and the entry of foreign objects. You can select quickly.

Specifically, the brightness difference is obtained by comparing the image of the monitoring camera with the reference image for each pixel, and the change pixel is extracted. As the reference image, an image from a predetermined time is used. The image may be one minute ago or one second ago. In the case of a fixed camera, the image may always be constant. Luminance difference by counting the number of changed pixels than the first threshold, when the number of changed pixels is equal to or more than the second threshold value, determines that there is a possibility of camera tampering. Further, it is checked whether or not there is a sudden change in the image. If there is a sudden change in the image, the luminance dispersion of the pixel is obtained, and when the luminance dispersion is smaller than the threshold value, it is determined that the monochromatic interference is present. Further, if the pixels that change brightly and the pixels that change darkly are approximately the same, it is determined that there is an interfering mixture change. When a camera disturbance is detected, a warning is issued by distinguishing whether a monochromatic disturbance or a mixed change disturbance has occurred.

  Next, the operation of each functional unit of the camera tampering detection method will be described with reference to FIG. One monitoring camera 1 or a plurality of monitoring cameras 1 may be provided. Check each camera 1 for camera interference. The image processing unit 2 performs image processing, statistical processing is performed on pixels having a luminance change, and monochromatic change interference and mixed change interference are detected. An image serving as a reference for comparison is held in the memory of the reference image storage unit 8. The image acquisition unit 3 acquires an image from the monitoring camera 1. The luminance difference calculation unit 4 compares the acquired image with the reference image to obtain a luminance difference for each pixel. The change pixel extraction unit 5 extracts change pixels having a luminance difference equal to or greater than the first threshold. The sudden change detection unit 6 determines that there is a sudden change in the image when the number of pixels having a luminance change equal to or greater than a threshold value between frames is equal to or greater than a predetermined value. The camera interference determination unit 7 determines whether the camera interference is a monochromatic interference or a mixed change interference from the data of the changed pixels. A warning reporting unit 9 notifies that camera interference has been detected.

  Next, the state of image change will be described with reference to FIGS. (+) Indicates a pixel region that changes brighter than a threshold value compared to the reference image. That is, the pixel is brighter than (reference value + threshold). (-) Indicates an area of a pixel that has changed darker than a threshold value compared to the reference image. That is, the pixel is darker than (reference value−threshold). The threshold value on the + side and the threshold value on the-side may be the same or different. As shown in FIG. 1B, the whole area changes in a bright direction when the illumination and sunshine change brightly. As shown in FIG.1 (c), the whole changes to a dark direction when illumination and sunlight change darkly. As shown in FIGS. 1D and 1E, when a part changes in a bright direction and a part changes in a dark direction and the ratio is close to 1, this is a case where the scene has changed greatly. That is, this is a case where the orientation of the camera has changed. As shown in FIGS. 1 (f) and (g), when a part of the scene changes to a bright direction and a part of the scene changes to a dark direction, and the ratio is considerably different from 1, a part of the scene changes greatly. Is the case. That is, when a large foreign object has entered.

  Next, the processing procedure of the camera tampering detection method will be described with reference to FIG. If there is a possibility of camera interference, the type of change is judged and a warning is issued. In step 1, a reference image is acquired. In step 2, an image is acquired from the camera. In step 3, a luminance difference value in each pixel between the current acquired image and the reference image is obtained constantly (or at regular intervals). If the luminance difference value is equal to or greater than the first threshold value in step 4, the pixel is changed on the image. In step 5, the pixels that have changed are counted. Step 6 repeats until all pixels are examined.

  In step 7, the amount of change pixel is checked. If there is no change pixel equal to or greater than the second threshold, it is checked in step 8 whether a foreign object has entered. If no foreign object has entered, the process ends. If a foreign object has entered, a foreign object intrusion warning is issued. It is assumed that there is a possibility of camera interference or illumination change when a change pixel occurs for all the pixels at the second threshold value or more. If there is a change pixel equal to or greater than the second threshold value, it is checked in step 9 whether it is a sudden change. For example, it is confirmed whether a large change has occurred suddenly from the past several frames. If there is no sudden change, it is checked in step 8 whether a foreign object has entered. The method of checking for foreign object intrusion is the same as the conventional method. If there is a sudden change, in step 10 it is checked if it is a single color change. If it is a single color change, an alarm to that effect is issued. If it is not a single color change, it is checked in step 11 if it is a mixed change. If it is a mixed change, an alarm to that effect is issued. If there is a possibility of camera interference, the type of change is judged and a warning is issued. If it is not mixed change, notification of illumination change is given.

  Explain how to check for sudden changes. When a foreign object comes from a distance from the far side of the screen or crosses the front of the screen, even if a large change on the image is greater than or equal to the second threshold value, the change occurs smoothly. It is different from instantaneous changes such as camera interference and lighting changes. Therefore, it is determined whether or not it is a sudden change, and only in the case of a sudden change, there is a possibility of camera disturbance or illumination change. There are various criteria for judging whether or not this is a sudden change. As an example 1, a change pixel amount having a luminance change equal to or greater than the first threshold in the inter-frame difference is examined. When the amount of change in the inter-frame difference is greater than or equal to the second threshold, a sudden large change is assumed. In this case, another threshold value may be used as the threshold value. As an example 2, when the average luminance value of each frame is calculated in the past several frames, the variance value of the average luminance value weighted with respect to the past several frames is calculated, and the variance value is equal to or greater than a predetermined threshold value Let's say a sudden big change. As an example 3, when the change pixel amount of each frame is calculated in the past several frames, the variance value of the change pixel amount weighted with respect to the past few frames is calculated, and the variance value is equal to or greater than a predetermined threshold value Let's say a sudden big change.

  A method for detecting monochromatic change interference will be described. A variance value of luminance values in all pixels of the acquired image is obtained. It is assumed that a single color change occurs when the variance value is small. The fact that the variance is so small that it is substantially zero means that the luminance values are concentrated at a certain value. In actual camera obstruction, if the camera is disconnected, there may be a complete monochrome (black), but if you cover the camera with your hand, spray, or tape, all pixels will not be completely monochrome. Some brightness unevenness occurs. Also, I don't know how many colors will change due to automatic camera correction. When it is covered with a hand, it becomes dark at first, but as the hand approaches, it changes to whitish due to automatic correction of the camera. Therefore, by looking at the dispersion value, it can be determined whether the color is monochromatic regardless of the color, and even if there is some luminance unevenness, it can be regarded as a monochromatic change.

  A method for detecting mixing change interference will be described. The ratio of a brightly changing pixel whose luminance difference value is brightly changed with a first threshold value or more and a darkly changing pixel whose absolute value of the luminance difference value is darker than a first threshold value is compared. As shown in FIGS. 1D and 1E, it is assumed that a mixing change has occurred when there is little deviation in the ratio. For example, when the ratio of the number of pixels that change brightly to the number of pixels that change darkly is half or more of all the pixels, and the ratio between them is 0.5 to 2.0, it is determined that the change is mixed. This ratio is determined according to the monitoring situation. As shown in FIGS. 1B and 1C, when the image is generally brighter or darker, it is assumed that an illumination change or a sudden sunshine change has occurred. Camera changes are not considered for lighting changes or sudden sunshine changes. As shown in FIGS. 1 (f) and (g), when the change is neither a monochromatic change nor a mixing change, it is determined that a large foreign object has entered rather than camera interference.

  As described above, in the embodiment of the present invention, the camera tampering detection method is performed by comparing the image of the surveillance camera and the reference image to obtain the luminance difference for each pixel, and if the luminance difference value is equal to or greater than the first threshold value. When a change pixel occurs and the change pixel occurs at the second threshold value or more, if there is a sudden change, the type of camera interference is determined and a warning is issued. Therefore, the camera interference can be determined only by image processing.

  The camera tampering detection method of the present invention is optimal as a method for easily and reliably detecting camera tampering, such as covering the surveillance camera or changing the orientation of the surveillance camera, up to the disturbance mode.

It is a functional block diagram of the camera tampering detection method in the embodiment of the present invention. It is a flowchart which shows the process sequence of the camera tampering detection system in the Example of this invention. It is a conceptual diagram which shows the conventional camera tampering detection system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Surveillance camera, 2 ... Image processing part, 3 ... Image acquisition part, 4 ... Luminance difference calculation part, 5 ... Change pixel extraction part, 6 ... Sudden change detection part, 7... Camera interference determination unit, 8... Reference image storage unit, 9.

Claims (6)

A memory that stores a reference image, a luminance difference calculation unit that calculates a luminance difference for each pixel by comparing the image of the monitoring camera with the reference image, and a pixel having a luminance difference equal to or greater than a first threshold is extracted as a change pixel A change pixel extraction unit that counts, a sudden change detection unit that detects a sudden change in an image, and a camera interference determination unit that determines the presence or absence of camera interference based on the number of change pixels and the presence or absence of a sudden change in an image And a warning notification unit that issues a warning when camera interference is detected ,
The camera interference determination unit is configured to determine that there is a possibility of camera interference when the number of the change pixels is equal to or greater than a second threshold, and a monochromatic interference detection unit that detects a monochromatic interference that changes an image to a single color; A mixed change disturbance detection means for detecting a mixed change disturbance that changes the scene of the entire image;
The mixed change interference detection means counts the number of bright change pixels brighter than the luminance obtained by adding the first threshold to the luminance of the pixel of the reference image, and subtracts the first threshold from the luminance of the pixel of the reference image. Means for counting the number of dark change pixels that are darker than the brightness, and means for determining the presence or absence of mixing change interference based on the ratio between the number of bright change pixels and the number of dark change pixels. Camera tampering detection method. The monochromatic interference detection unit includes: a unit for obtaining luminance dispersion of all pixels; and a unit for determining that there is monochromatic interference when the luminance dispersion is substantially zero . Camera tampering detection method. A memory that stores a reference image, a luminance difference calculation unit that calculates a luminance difference for each pixel by comparing the image of the monitoring camera with the reference image, and a pixel having a luminance difference equal to or greater than a first threshold is extracted as a change pixel A change pixel extraction unit that counts, a sudden change detection unit that detects a sudden change in an image, and a camera interference determination unit that determines the presence or absence of camera interference based on the number of change pixels and the presence or absence of a sudden change in an image And a warning notification unit that issues a warning when camera interference is detected,
The sudden change detection unit includes means for determining that there is a sudden change in an image when the number of pixels having a luminance change greater than or equal to the first threshold value between frames is greater than or equal to the second threshold value . Luke camera interference detection method. A luminance difference calculation step of calculating a luminance difference for each pixel by comparing the image of the monitoring camera and a reference image stored in the memory, and a changed pixel that extracts and counts a pixel having a luminance difference equal to or greater than a first threshold as a changed pixel An extraction step, a sudden change detection step for detecting a sudden change in the image, a camera disturbance determination step for determining the presence or absence of camera interference based on the number of the change pixels and the presence or absence of the sudden change in the image, and a camera interference A warning issuing process for issuing a warning when an error is detected,
The sudden change detecting step, wherein Rukoto the number of pixels with luminance change than the first threshold value between frames having a step of determining that there is a sudden change in the image when it is more than the second threshold value A camera tampering detection method in the camera tampering detection method . 5. The camera tampering detection method in the camera tampering detection method according to claim 4 , wherein the luminance variance of all pixels is obtained and it is determined that the monochromatic interference is present when the luminance variance is substantially zero . The number of bright change pixels brighter than the luminance obtained by adding the first threshold to the luminance of the pixels of the reference image is counted, and the number of dark change pixels darker than the luminance obtained by subtracting the first threshold from the luminance of the pixels of the reference image 5. The camera tampering detection method in the camera tampering detection method according to claim 4 , wherein the presence / absence of mixing change tampering is determined based on a ratio between the number of bright change pixels and the number of dark change pixels. .

Images