JP2011061651A - Suspicious object detection system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem wherein, when there is a person watching a monitor, a person for removing a suspicious object is available immediately but if there is no monitoring person, even if there is any suspicious object on a monitor display image being monitored, the suspicious object can not be discovered soon, or if a plurality of two-dimensional (2D) imaging apparatuses are installed, since the monitoring person has to confirm images as many as the installed apparatuses, confirmation may takes more time. <P>SOLUTION: A suspect A suspicious object detection system is equipped with: a 2D imaging apparatus for imaging the inside of a monitoring view field; a storage device for recording a plurality of reference images having different levels of lightness and an image captured by the 2D imaging apparatus; an image processing device for comparing lightness of the image captured by the 2D imaging apparatus with a reference image having lightness closest thereto to detect a suspicious object; and a warning device which raises an alarm. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a suspicious object detection system that monitors with a monitoring image taken by a two-dimensional imaging device such as a CCD camera, and compares a plurality of pre-registered reference images with a monitoring image and detects a suspicious object according to an image processing result. It is related to the system.

A conventional suspicious object detection system detects a suspicious object using a two-dimensional imaging device such as an analog fixed camera. In such a suspicious object detection system, a monitoring image captured and acquired by a two-dimensional imaging device is not recorded in a memory, and a monitoring person periodically looks at a monitoring image on a monitor to detect a suspicious object. The presence or absence was confirmed.
A conventional suspicious object detection system will be described with reference to FIG. 1. FIG. 1 is a diagram for explaining an example of a conventional suspicious object detection system. 101 is a monitor display screen, 103 is an eye schematically showing that the supervisor is monitoring, and 104 to 106 are images on the monitor display screen 101. That is, 104 is a proposal, 105 is a suspicious object on the monitor display screen 101, and 106 is a person to be removed on the monitor display screen 101.

  FIG. 1 is a diagram for explaining a suspicious object detection system in which a top portion of a dam, which is also a passage, is photographed by a two-dimensional imaging device within a monitoring visual field and monitored in a monitoring room or the like. FIG. 1A shows a case where the monitor is monitoring the monitor, and the monitor looks at the monitor display screen 101 and finds that the suspicious object 105 exists on the proposal 104. This indicates that the person who asks the person 106 to remove the suspicious object 105 is requested. However, as shown in FIG. 1B, the suspicious object 105 is displayed on the monitor display screen 101. However, since there is no person to monitor, it is not known that the suspicious object 105 exists on the bank top 104. Therefore, even if there is a suspicious object 105, it is left as it is. If the suspicious object 105 is a time bomb, the presence of the dam cannot be confirmed until the monitor next looks at the monitor display screen 101, so that the dam may be destroyed and serious damage may occur.

JP 2008-341

If there is a monitor as in the conventional example, you can request a person to remove it immediately, but if there is no monitor, even if there is a suspicious object on the monitor display image to be monitored, it will be detected immediately Can not do it. In addition, when a plurality of two-dimensional imaging devices are installed, there is a problem that it takes time for confirmation because the supervisor must perform confirmation for the number of the two-dimensional imaging devices.
Furthermore, in the conventional suspicious object detection system, the monitoring image is not recorded, and the image captured by the two-dimensional imaging device is displayed on the monitor and confirmed by the observer. This necessitated regular intervention by the observer. Therefore, if the number of installed two-dimensional imaging devices is large, it takes a long time for the confirmation work, and there is a possibility that the suspicious object may be left until the supervisor finds the suspicious object.
Therefore, it is necessary to automatically identify a suspicious object using an image processing apparatus. Patent Document 1 describes a technique for detecting a suspicious object left in each machine in various game halls such as a pachinko machine and a pachislot machine by image comparison. However, a two-dimensional imaging device is photographed under a certain degree of brightness, such as indoors, and the obtained video is subjected to image processing to identify a suspicious object. Therefore, it is used outdoors like dam monitoring, and even in the daytime, the brightness within the monitoring field of view is affected by changes in climate such as sunny, rainy, cloudy, and the influence of shadows caused by clouds and trees. There is no support for extreme changes.
In view of the above problems, the object of the present invention is to monitor a suspicious object even when the brightness in the monitoring visual field changes extremely, and to record a monitoring image and to check the image visually by the observer. The object is to provide a suspicious object detection system that can be monitored even when the number of installed two-dimensional imaging devices is large.

In order to solve the above problems, a suspicious object detection system according to the present invention records a two-dimensional imaging device that captures an image within a monitoring visual field, a plurality of reference images having different brightness, and an image captured by the two-dimensional imaging device. A storage device, an image processing device that detects a suspicious object by comparing the brightness of an image captured by the two-dimensional imaging device with a reference image having the closest brightness, and an alarm device that issues an alarm. Is.
Preferably, the storage device of the suspicious object detection system of the invention further records a processing result of the image processing unit.

  According to the present invention, it is possible to select an appropriate reference image according to the environment by comparing an image captured by the two-dimensional imaging device in the monitoring visual field with a reference image closest to the brightness of the image. It is possible to detect only suspicious objects more accurately. In addition, since the reference image is selected based on the brightness, it is possible to prevent a shadow or reflection due to a change in sunlight from being judged as a suspicious object. In addition, the alarm device and recording device can be used together to output an alarm output when a suspicious object is detected, display a suspicious object image, record a monitoring image and an image processing result image, and quickly check the suspicious object. It is.

It is a figure for demonstrating an example of the conventional suspicious object detection system. It is a systematic diagram for demonstrating one Example of the suspicious object detection system of this invention. It is a block diagram which shows the structure of one Example of the suspicious object detection system of this invention. It is the flowchart and reference drawing for demonstrating the processing operation of one Example of the suspicious object detection system of this invention. It is a figure for demonstrating setting a mask area | region when the shift | offset | difference arises in the angle of view of a monitoring image and a reference image.

The suspicious object detection system according to the present invention is an image obtained by photographing the inside of a monitoring visual field with a two-dimensional imaging device and photographing the inside of the monitoring visual field with an image processing device from a plurality of reference images having different brightnesses recorded in advance in a storage device. A comparison is made with a reference image having the closest brightness, a suspicious object is detected, and an alarm is issued by an alarm device.
Preferably, in the suspicious object detection system according to the invention, the storage device further records a processing result of the image processing device.
Preferably, in the suspicious object detection system of the above invention, in order to detect the suspicious object more reliably, it is necessary to cope with an environmental change. Each time a suspicious object is confirmed, it is reflected in the update of the reference image based on the result of image processing, and the latest monitoring environment is maintained, so that it is possible to cope with the brightness change and disturbance change of the monitoring image.

Preferably, in the suspicious object detection system according to the present invention, a pattern (a plurality of patterns) is set in the reference image as an index of brightness of the selected reference image. Compare the brightness judgment pattern registered in advance with the brightness of the image in the monitoring field captured by the 2D imaging device, select the reference image closest to the result, and change the brightness or disturbance of the monitoring image It corresponds to change.
More preferably, in the above suspicious object detection system of the present invention, by registering a positioning pattern of the reference image, even if the angle of view of the monitoring image taken by the two-dimensional imaging device is shifted, it is determined as the same image by pattern matching. This corresponds to a video shift.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In addition, in the description of each drawing including FIG. 1 already described, the same reference numerals are given to the constituent elements having a common function, and the overlapping description is avoided as much as possible.

  A suspicious object detection system according to an embodiment of the present invention will be described with reference to FIGS. FIG. 2 is a system diagram for explaining an embodiment of the suspicious object detection system of the present invention. Also. FIG. 3 is a block diagram showing the configuration of an embodiment of the suspicious object detection system of the present invention. 201 is a two-dimensional imaging device such as a CCD camera, 202 is an image processing device, 203 is a monitor, 204 is a storage device, and 205 is an alarm device. In addition, 301 is a two-dimensional imaging device such as a CCD camera, 302 is an image processing device, 303 is a monitor, 304 is a storage device, 304 is an alarm device, 306 is a monitoring field of view, and 307 is a monitor that needs to detect the presence or absence of a suspicious object. The object 308 is a suspicious object.

In FIG. 2, the suspicious object detection system of the present invention includes at least a two-dimensional imaging device 201, an image processing device 202, a monitor 203, a storage device 204, and an alarm device 205. The connection between the camera 201 and the image processing apparatus 202, and the transmission lines between the image processing apparatus 202 and the monitor 203, the storage device 204, and the alarm device 205 may be wireless, wired, or a combination thereof. The transmission lines may be a coaxial cable, a LAN (Local Area Network), a WAN (Wide Area Network), or a combination thereof.
The two-dimensional imaging device 201 is, for example, a CCD camera, a network camera, or a high sensitivity camera.
The image captured by the two-dimensional imaging device 201 is output as a video signal, and the image is recorded in the storage device 203 as it is without being compressed.

The suspicious object detection system configuration of the present invention will be further described with reference to FIG.
In the storage device 304, reference images with different brightnesses extracted based on images in the monitoring visual field 306 are recorded in advance. An example of the recording procedure will be described below.
The monitor 303 is provided with an HMI (Human Machine Interface), and the monitor can perform an operation such as a GUI (Graphical User Interface) operation while viewing the display screen.

In FIG. 3, the monitor operates the monitor 303 to set the reference image registration mode. Already, in the storage device 304, the two-dimensional imaging device 301 has captured the inside of the visual field range 306 with different brightness, and these images are recorded. The different brightness can be obtained, for example, by photographing the visual field range 306 one or more times 24 hours a day and night. A large tool may be set so that a shadow is artificially generated in the monitoring visual field 306.
As described above, the monitor classifies the brightness from the images having different brightness accumulated in the storage device 304, sets a reference image for each brightness, associates the brightness with the class of brightness, and stores it. Register with device 304. A threshold value for binarization processing to be described later is registered in the storage device 304 in association with the reference image.
Before determining the above-mentioned reference image, confirm the matching (presence / absence of field angle deviation) between the monitoring image and the selected reference image, and perform positioning correction (field angle adjustment). It is necessary to define a judgment pattern. A predetermined range is registered in the storage device 304 or the image processing device 302 as an alignment pattern. As the positioning determination pattern, a part with little change in brightness or video is selected and registered.

In order to classify the brightness in the image captured by the two-dimensional imaging device 301 and the reference image, for example, a partial region of the image is defined as a brightness determination pattern, and the brightness of each pixel in the pattern is determined. A representative value is obtained, and classification is performed based on the obtained representative value. The representative value may be an average value or a median value, for example.
Further, for example, a representative value of a value excluding the maximum value and the minimum value of luminance in the determination pattern may be used.
These records may be recorded by the storage device 304 or the image processing device 302.

Preferably, the threshold value for binarization may be different for each brightness. Further, as the brightness increases, the threshold value for each brightness class may be gradually decreased, or conversely, the threshold value may be gradually increased.
Preferably, the threshold value of the predetermined brightness class may be set to the minimum value, and the threshold value may be gradually increased in the brightness classes before and after that. Furthermore, the threshold value of the predetermined brightness class may be set to the maximum value, and the threshold value may be gradually decreased in the brightness classes before and after that.
More preferably, a program for automatically executing the above processing may be installed in the monitor 303 or the image processing apparatus 302, and the reference image may be automatically registered in accordance with the program execution instruction of the supervisor.

Next, the operation of the suspicious object detection system will be described with reference to FIG.
In FIG. 3, the two-dimensional imaging device 301 captures the inside of the monitoring visual field 306 and outputs the acquired image to the image processing device 302. The image processing device 302 outputs the input image to the monitor 303 and the recording device 304. The monitor 303 displays the input image on the display screen. The storage device 304 records the input image.
Further, the image processing device 202 detects the brightness of a predetermined area of the image input from the two-dimensional imaging device 301. The detected brightness is classified in a predetermined brightness (luminance value) range so that the brightness can be classified. Further, the classified brightness range corresponds to the brightness classification of the reference image registered in advance in the storage device 304.

The two-dimensional imaging unit 301 captures the inside of the monitoring visual field 306 and outputs the captured image to the image processing apparatus 302 as a video signal. The image processing device 302 detects the brightness of the input image, checks which brightness class the detected brightness belongs to, and stores the reference image of the corresponding brightness class from the storage device 304. read out.
The image processing device 302 binarizes the input image (monitoring image) and the image read from the storage device 304, and obtains a difference for each pixel. Then, a pixel region having a difference value equal to or greater than a predetermined threshold is extracted. The extracted pixel area detects a cluster (pixel area) of pixels having a predetermined size or more as a suspicious object. In other words, the difference image between the monitoring image and the reference image usually contains noise due to disturbance or video change in addition to the object to be detected. To extract.
More preferably, the image processing device 302 performs expansion / contraction processing, numbering, noise component removal processing, and the like on the extracted pixel region, and then sets a pixel block (pixel region) having a predetermined size or more as a suspicious object. Detect.

When the suspicious object is detected as described above, the image processing apparatus 302 transmits a signal notifying the alarm device 305 that the suspicious object has been detected. The alarm device 305 outputs an alarm when receiving a notification signal.
Further, the image processing device 302 transmits a signal notifying the alarm device 305 that a suspicious object has been detected, and outputs a detection image after image processing to the monitor 303 to display it on the display screen of the monitor 303.
Further, the image processing apparatus 302 outputs the detected image of the suspicious object and the detected image after the image processing to the storage device 304 and causes the storage device 304 to record the detected image.

The processing operation of the suspicious object detection system of the present invention will be described with reference to FIGS. FIG. 4 is a flowchart and a reference diagram for explaining the processing operation of the embodiment of the suspicious object detection system of the present invention. FIG. 4A is a flowchart of the processing operation. FIG. 4B shows the monitoring image 401 captured in step S1. FIG. 4C shows reference images T-1, T-2, T-3, and T-4 registered in the storage device 304, respectively. FIG. 4D is the difference image 441 in step S6. FIG. 4E shows the image processed image 451 detected in step S7. FIG. 5 is a diagram for describing setting of a mask area when a deviation occurs in the angle of view between the monitoring image 501 and the reference image 502.
In the flowchart of FIG. 4A, the processing operation is executed by the image processing apparatus 302 unless otherwise specified.

When the suspicious object detection system is operated, in step S1, the monitoring image 401 input from the two-dimensional imaging device 301 is first captured. The monitoring image 401 includes a suspicious object 405, a brightness determination pattern 407, and a positioning determination pattern 408.
In step S 2, the monitoring image is recorded in the storage device 304.
In step S3, the brightness of the brightness determination pattern 407 portion of the monitoring image is calculated.
In step S4, the reference image T-1 closest to the brightness of the monitoring image 401 is selected from the result, and is read from the storage device 304.

In step S5, an image processing area is set.
That is, in order to determine the presence or absence of a suspicious object, the reference image (plurality) is registered in advance as described above. The reference image may be various, such as a general image that seems to have the highest frequency, an image affected by changes in sunshine, a bright image, a dark image (an image with a dark cloud, etc.). In the brightness determination pattern, the reference brightness is determined in order to select the reference image based on the brightness of the monitoring image, and the location is registered as the brightness determination pattern. Further, in this case, a part that is easily affected by brightness (an area such as a part exposed to sunlight or a part having a luminance change) is selected.

Next, the consistency (presence / absence of field angle deviation) between the monitoring image and the selected reference image is confirmed. For this reason, using the positioning determination pattern, the portion where the image is shifted and is not reflected in the reference image is masked, and the portion other than the masked portion (mask region 508) is set as the image processing region (see FIG. 5). If there is less than a predetermined amount (number of pixels), it is assumed that there is no deviation.
Further, there may be a case where the region of the positioning determination pattern is not shown on the monitoring image due to the angle of view deviation. In this case, the detection process of the positioning determination pattern is performed a predetermined number of times, and if it cannot be detected, the image processing is not performed and an error in the angle of view is returned.

In step S6, image processing is performed on the monitoring image and the reference image. In this embodiment, the difference process and the binarization process are performed, but the present invention is not limited to this process.
In step S7, the presence or absence of a suspicious object is detected using the result of step S6 as an image processed image. If there is no suspicious object, the process proceeds to step S10. If a suspicious object is detected, the process proceeds to step S8.
In step S8, the image processed image is recorded in the storage device 304.
Subsequently, in step S9, data is output to the alarm device 305 and the monitor 303 in order to reflect the determination result. In this way, when there is a suspicious object, the supervisor is contacted.
In step S10, the reference image is updated in order to record the environmental change of the current monitoring video. The reference image to be updated is an image selected before executing image processing. Thereby, since it is possible to sequentially cope with a minute environmental change, it is possible to select an appropriate reference image and to detect a suspicious object more accurately.
Furthermore, in step S11, if there are other monitoring locations and there are monitoring images input from other two-dimensional imaging devices, the process returns to step S1 and the same processing is repeated.

According to the above-described embodiment, an appropriate reference image according to the environment is selected by comparing an image captured by the two-dimensional imaging device in the monitoring visual field with the reference image closest to the brightness of the image. It is possible to detect only suspicious objects more accurately. Further, since the reference image is selected based on the brightness, it is possible to prevent a shadow or reflection due to a change in sunlight from being judged as a suspicious object. In addition, the alarm device and recording device can be used together to output an alarm output when a suspicious object is detected, display a suspicious object image, record a monitoring image and an image processing result image, and quickly check the suspicious object. It is.
As a result, when a suspicious object is detected in the image processing image, data can be output to an output unit such as an alarm device and a monitor, and a monitor can be contacted. Thereby, stable detection can be performed without being influenced by the environmental change of the monitoring image, and the monitoring person can quickly confirm and remove the suspicious object as necessary.

  101: Monitor display screen, 103: Eyes indicating that the supervisor is monitoring, 104: Proposition, 105: Suspicious object, 106: Person to remove, 201: Two-dimensional imaging device, 202: Image processing Device 203: monitor 204: storage device 205: alarm device 301: two-dimensional imaging device 302: image processing device 303: monitor 304: storage device 305: alarm device 306: monitoring field of view 307: Monitored object, 308: Suspicious object.

Claims (1)

  A two-dimensional imaging device that captures the inside of the monitoring field of view, a storage device that records a plurality of reference images having different brightness and an image captured by the two-dimensional imaging device, and the brightness of the image captured by the two-dimensional imaging device is the highest. A suspicious object detection system comprising: an image processing device that detects a suspicious object by comparing with a reference image having near brightness; and an alarm device that issues an alarm.

Images