JP2009294974A - Intruder detection device by image processing - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a simple-configuration and high-reliability intruder detection device by image processing. <P>SOLUTION: This intruder detection device by the image processing has: a monitoring camera 1 imaging a monitoring target area; and an image processing part 2 detecting an intrusion object based on a result obtained by processing an input image imaged by the monitoring camera 1. The image processing part 2 has: a storage part 21; a processing setting part 22 setting at least an intrusion detection confirmation time; an image input part 23 for inputting the input image; an image change detection part 24 detecting an area wherein a change occurs on an image by comparing the input image and a previously imaged background image; an intrusion object candidate detection part 25 comparing a geometric characteristic amount of the area wherein the change occurs on the image and a preset characteristic amount range of the intrusion object to detect an intrusion object candidate; an intrusion object detection part 26 detecting the intrusion object candidate as the intrusion object when the intrusion object candidate is detected continuously for the intrusion detection confirmation time or above; and a result output part 27 outputting a detection result. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an intruder detection device by image processing for detecting an intruder into a monitoring target area such as a prohibited entry point in a facility or road, and in particular, a monitoring camera for imaging the monitoring target area and its surroundings is installed, The present invention relates to an intruder detection device based on image processing that detects a person or the like who enters a prohibited entry area by processing an image captured by the monitoring camera.

  2. Description of the Related Art Conventionally, intruder detection devices using image processing for outdoor facilities are known. For example, what is disclosed in Patent Document 1 is a region (partial region) in which an image obtained by capturing a monitoring target region with a monitoring camera is compared with a background image captured in advance and a change is generated on the image. ), And the presence or absence of an intruder is determined by comparing the geometric feature amount of the partial region with a preset feature amount range of the intruder, and the position of the intruder in the image is specified. . The thing of this patent document 1 has the advantage that an intruder can be detected more accurately using the existing surveillance camera.

  In addition, what is disclosed in Patent Document 2 uses a stereo camera in which two cameras are installed in parallel as a surveillance camera in order to prevent misinformation due to external light, and takes the difference between the left and right images to determine the left and right. A difference image is obtained, and the left and right difference image is compared with the set background image stored in advance to extract a portion where a change has occurred on the image. The part where the change has occurred is determined as an intruder. This utilizes the fact that the size of the changed portion due to external light is in the error range in the left-right difference image.

  In addition, what is disclosed in Patent Document 3 is provided with an imaging unit and a laser distance measuring unit, in order to eliminate the possibility of misreporting light / shadow patterns and small animals as intruders. In addition to comparing the image with the reference image, the intruder is determined by comparing the distance to the object measured by the laser ranging unit and the reference distance.

JP 2006-252248 A JP 2007-124269 A JP 2002-208073 A

  However, the prior art described in Patent Document 1 is configured to detect an intruder using only image information, and may be caused by temporary changes in sunlight, sudden fluctuations in vegetation, noise in image signals, and the like. There has been a problem that a false alarm may occur when a change occurs on an image. In particular, outdoor facilities are subject to monitoring, including “stopped person (hereinafter referred to as“ stopped person ”)”, “pedestrian”, “running person (hereinafter referred to as“ runner ”)”, “bicycle”, “automobile”, “ When a plurality of objects such as “forklifts” are set as objects of intruders, the feature amount range of the invaders to be set is widened, so that there is a possibility that false alarms are generated.

  Moreover, in the prior art described in Patent Document 2, two cameras are required as an input device, which is expensive, and installation work is required such that two cameras need to be arranged at a predetermined interval. There was a problem that it could be complicated.

  Moreover, in the prior art described in the above-mentioned Patent Document 3, it is necessary to install a laser distance measuring unit for each monitoring location, so that there is a problem of cost, and the place where the monitoring device can be installed depends on the laser distance measuring unit. Since it is limited to a place where there is no obstruction that hinders measurement, there may be a problem that installation becomes difficult. Furthermore, since the object measured by the imaging unit is configured to detect the distance to the object in the laser ranging unit, it is necessary to align the imaging unit and the laser ranging unit. In addition, there is a problem that adjustment when installing the apparatus is complicated.

  In view of the above, an object of the present invention is to provide an intruder detection apparatus using image processing with a simple configuration and high reliability.

  An intruder detection apparatus using image processing according to the first invention for solving the above-described problem is an intruder based on a result of processing an input image captured by the imaging unit and an imaging unit that images a monitoring target region An image processing means for detecting an object, wherein the image processing means is a storage means, a processing setting means for setting at least an intrusion detection confirmation time, an image input means for inputting the input image, and an image picked up in advance with the input image. An image change detecting means for detecting a region where a change has occurred on the image by comparing with a background image, and a geometric feature amount of the region having a change on the image and an intruder set in advance Intruder candidate detecting means for detecting an intruder candidate by comparing with a feature amount range of the intruder, and detecting the intruder candidate as an intruder when the intruder candidate is continuously detected for the intrusion detection confirmation time or longer. Intrusion Detecting means, and characterized in that it has a result output means for outputting a detection result.

  According to the intruder detection device by the image processing according to the present invention described above, in the intruder detection device that detects an intruder into the monitoring target area of the outdoor facility, the temporary sunshine, the vegetation caused by the sudden wind False alarms due to shaking, noise in the image signal, etc. can be suppressed, and a highly reliable intruder detection result can be obtained.

  In addition, it is not necessary to use a special device such as a stereo camera as an imaging means, and it is possible to detect an intruder into the monitoring target area with high accuracy while suppressing the introduction cost of the monitoring system using a general imaging means. Can do.

  Furthermore, since it is not necessary to use the image pickup means and another sensor device in combination, the apparatus can be easily installed.

  Details of embodiments of the present invention will be described in the following examples.

  1 is a block diagram illustrating a schematic configuration of an intruder detection apparatus according to the present embodiment, FIG. 2 is a flowchart illustrating processing in an image processing unit according to the present embodiment, and FIG. 3 is an image captured by a monitoring camera according to the present embodiment. FIG. 4 is an explanatory diagram illustrating another example of an image captured by the surveillance camera in the present embodiment, and FIG. 5A is an explanatory diagram illustrating an example in which the input image is divided into small regions. FIG. 5B is an explanatory diagram illustrating an example in which an input image is divided into a change small region and a background small region, and FIG. 5C is an explanatory diagram illustrating a setting example of a change detection small region.

  The intruder detection device according to the present embodiment is a device that detects an intruder into a monitoring target area such as a prohibited entry point in a facility or road, and as shown in FIG. The image processing unit 2 is configured as an image processing unit that analyzes an image captured by the monitoring camera 1.

  The surveillance camera 1 is installed at a position where a monitoring target area such as a prohibited entry point on a facility or road can be seen over, and continuously images the monitoring target area and its peripheral area, and the captured image is used as image data. It is configured to transmit to the processing unit 2.

  On the other hand, the image processing unit 2 includes a storage unit 21 as a storage unit, a process setting unit 22 as a process setting unit, an image input unit 23 as an image input unit, an image change detection unit 24 as an image change detection unit, an intruder The intruder candidate detection unit 25 as a candidate detection unit, the intrusion detection unit 26 as an intrusion detection unit, and a result output unit 27 as a result output unit.

  The storage unit 21 is a part in which various data is stored. For example, background image data obtained by capturing an image of the monitoring target area in advance is stored.

  The process setting unit 22 is a part for setting parameters (for example, an intruder feature amount range, an intrusion detection confirmation time, etc.) necessary for the determination of intruder detection. The set process parameters (process variables) are stored in the storage unit 21. Stored in In the present embodiment, the intrusion detection confirmation time is a time set in advance to determine the presence or absence of an intruder, and is appropriately set according to various conditions.

  The image input unit 23 is a part that captures image data transmitted from the surveillance camera 1 at a predetermined time interval, and the captured image data is stored in the storage unit 21 as input image data.

  The image change detection unit 24 is a part that detects an area where a change has occurred by comparing the brightness of two images and the like, and includes input image data, background image data, and processing parameters (in this embodiment, a correlation value described later). A threshold value for determining whether the temporal change corresponds to “occurrence of change” or “little change from the background”) from the storage unit 21, and divides the image into appropriate small regions; Calculates the correlation between the background image data corresponding to each small area and the input image data, and detects the change from the small areas set in the image based on the temporal change of the correlation value. The changed small region and the background candidate small region with little change in the background are detected, and the obtained result is stored in the storage unit 21 as small region data.

  The intruder candidate detection unit 25 takes out the small area data and the processing parameters (in this embodiment, the feature amount range corresponding to the intruder) from the storage unit 21, and combines the adjacent small change areas to obtain the change detection partial area. The change detection partial region is a portion that is detected as an intruder candidate when the geometric feature amount of the change detection partial region is within the feature amount range corresponding to the intruder set in advance. The obtained result is stored in the storage unit 21 as intruder candidate data.

  The intruder detection unit 26 extracts the intruder candidate data and the processing parameters (intrusion detection confirmation time in this embodiment) from the storage unit 21, and the area detected as the intruder candidate continues for a preset intrusion detection confirmation time or longer. This is a part for determining whether or not an intruder candidate is detected as an intruder continuously by determining whether or not the intruder has been detected. The obtained result is stored in the storage unit 21 as intruder data.

  The result output unit 27 retrieves the input image data and the intruder data from the storage unit 21, and the input image data, presence / absence of intruder detection obtained based on the intruder data, the position of the intruder candidate on the image, and the intruder This is a part for outputting detection result data composed of geometric data representing the type (for example, person, vehicle, etc.).

  The flow of processing by the image processing unit of this embodiment will be described below with reference to FIGS.

  As shown in FIG. 2, in this embodiment, first, image data as shown in FIGS. 3 and 4 continuously transmitted from the monitoring camera 1 in the image input unit 23 is fetched as input image data and stored in the storage unit 21. Store (step S1).

  Thereafter, in the image change detection unit 24, the input image data, the background image data stored in the storage unit in advance, and the processing variables are taken out from the storage unit 21, and for example, as shown in FIG. It is obtained by dividing the image into appropriate small regions 31 and comparing the input image 3 with the background image based on the processing variables to detect the change small region 31a and the background candidate small region 31b as shown in FIG. 5B. The result is stored in the storage unit 21 as small area data (step S2).

  Subsequently, the intruder candidate detection unit 25 takes out the small region data and the feature amount range corresponding to the intruder from the storage unit 21, and changes by combining the adjacent small change regions 31a as shown in FIG. When the detection partial area 31c is determined, and the geometric feature amount of the change detection partial area 31c is within the preset feature amount range corresponding to the intruder, the intruder candidate 4 is present in the change detection partial area 31c. The detection result is stored in the storage unit 21 as intruder candidate data (step S3).

  Subsequently, the intruder detection unit 26 extracts the intruder candidate data and the processing variable from the storage unit 21, and determines whether or not the intruder candidate 4 is continuously detected for the intrusion detection confirmation time or more. 3, the intruder candidate 4 is regarded as the intruder 5 when the intruder candidate 4 is continuously detected on the image for at least the intrusion detection confirmation time (for example, t = n or more in FIG. 3). This is stored in the storage unit 21 as intruder data (step S4).

  Thereafter, in the result output unit 27, the image data and the intruder data are taken out from the storage unit 21, the input image data, the presence / absence of intruder detection obtained based on the intruder data, the position of the intruder candidate 4 on the image, And the data comprised by geometric data etc. are output as a detection result (step S5).

  As described above, when a plurality of types such as “stopper”, “pedestrian”, “runner”, “bicycle”, “automobile”, “forklift”, etc. are set as targets of the intruder to be detected. Because the range of the feature value corresponding to the intruder becomes wide, the change detection partial area detected by the change on the image due to the temporary sunshine change, the shaking of the plant due to sudden wind, the noise of the image signal, etc. There is a greater possibility that the feature amount of the item matches the “range of feature amount corresponding to the intruder”.

  Therefore, in the device described in Patent Document 1, the feature amount of the change detection partial area detected from a change in the image due to a temporary change in sunshine, a sudden swaying of the vegetation, noise in the image signal, etc. However, when it is within the feature amount range corresponding to the intruder, intruder detection is generated, and this intruder detection may be a false detection.

  On the other hand, according to the intruder detection device according to the present embodiment, when the intruder detection unit 26 detects the intruder candidate, whether or not the intruder candidate appears on the image continuously for a predetermined time. And the presence or absence of an intruder based on the result of the determination, it is possible to suppress false alarms due to temporary changes in sunshine, sudden fluctuations in vegetation, image signal noise, etc. High intruder detection results can be obtained.

  In addition, unlike the techniques described in Patent Documents 2 and 3 described above, a special configuration is not required for the input device, and thus an intruder detection device can be configured using a general monitoring camera. Therefore, the existing surveillance camera installed in the outdoor facility can be used, and the introduction cost of the surveillance system can be kept low.

  Furthermore, as described in Patent Document 3, a configuration in which an intruder is detected by analyzing an image captured by the surveillance camera is employed without using a configuration in which the surveillance camera and another sensor device are used in combination. Therefore, it is not necessary to make complicated adjustments such as positioning the surveillance camera with another sensor device, and adjustment at the time of device installation is facilitated.

  In addition, this invention is not limited to the Example mentioned above, It cannot be overemphasized that a various change is possible in the range which does not deviate from the meaning of this invention.

  The present invention relates to an intruder detection device by image processing for detecting an intruder in a facility or road entry prohibition, and the like, and in particular, a surveillance camera for monitoring an entry prohibition site in an outdoor facility is installed, and the surveillance camera captures the image. It can be applied to an intruder detection device based on image processing that detects a person or the like entering an entry-prohibited location in the image by processing the processed image.

It is a block diagram which shows schematic structure of the intruder detection apparatus which concerns on the Example of this invention. It is a flowchart which shows the flow of a process in the image process part of the Example of this invention. It is explanatory drawing which shows the example of the image imaged with the surveillance camera in the implementation low of this invention. It is explanatory drawing which shows the example of the other image imaged with the surveillance camera in the implementation low of this invention. FIG. 5A is an explanatory diagram showing an example of dividing an input image into small regions, FIG. 5B is an explanatory diagram showing an example of setting a change small region and a background candidate small region, and FIG. 5C is a change detection. It is explanatory drawing which shows the example of a setting of a partial area.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Surveillance camera 2 Image processing part 3 Input image 21 Memory | storage part 22 Process setting part 23 Image input part 24 Image change detection part 25 Intruder detection part 26 Intrusion detection part 27 Result output part 31 Small area 31a Change small area 31b Background Candidate small area 31c Change detection partial area

Claims (1)

An imaging unit that images a monitoring target region; and an image processing unit that detects an intruder based on a result of processing an input image captured by the imaging unit;
The image processing means is a storage means, a process setting means for setting at least an intrusion detection confirmation time, an image input means for inputting the input image, an image obtained by comparing the input image with a background image captured in advance. An image change detection means for detecting a region where a change has occurred on the top, and a geometric feature amount of the region where the change has occurred on the image is compared with a preset feature amount range of the intruder, so that an intruder candidate An intruder candidate detecting means for detecting the intruder candidate, an intruder detecting means for detecting the intruder candidate as an intruder when the intruder candidate is continuously detected for the intrusion detection confirmation time or more, and a detection result is output. An intruder detection apparatus using image processing, characterized by comprising a result output means.

Images