JP5045449B2 - Environment-adaptive intruder detection device by image processing - Google Patents

Description

  The present invention relates to an apparatus for detecting an intruder into a facility or road entry prohibition place, and in particular, by installing a surveillance camera for monitoring the entry prohibition place and processing the captured image, The present invention relates to an intruder detection device that detects a person or the like who enters a prohibited entry area.

  Conventionally, an intruder detection device composed of a monitoring camera and an image processing unit has been used to detect intrusions into restricted areas by analyzing images of surveillance cameras installed in places where entry is prohibited such as facilities and roads. Yes (see, for example, Patent Documents 1 to 3). In the intruder detection devices disclosed in Patent Documents 1 to 3, the surveillance camera is installed at a position overlooking the prohibited entry area, images the prohibited entry area and its peripheral part, and transmits the image to the image processing unit. Then, the image processing unit inputs the image of the surveillance camera, divides the image into appropriate small areas, and obtains a background image acquired in advance for each small area and an image of the corresponding small area in the input image. The change detection area is determined by combining the adjacent change small areas and detecting the change small area and the background candidate small area in the image based on the temporal change. When the geometric feature amount of the intruder is within the feature amount range corresponding to the intruder set in advance, it is determined that the intruder exists in the change detection area, and the intruder is detected and the position in the image Is configured to identify.

JP 2006-252248 A JP 2007-233919 A JP 2007-041616 A

  In Patent Documents 1 to 3, an input image is compared with a background image set in advance, and an intruder that has entered a prohibited entry area or the like and is captured by a surveillance camera is detected as a difference from the background image. Thus, the intruder is detected.

  However, in such an intruder detection device, if the place to be imaged by the surveillance camera, that is, the place to be monitored is a place where there is little environmental change, for example, indoors, Although it is considered that the environment (for example, illuminance) changes little when the background image compared with the input image is captured, the location of the monitoring target changes, for example, outdoors as time passes. If the environment is subject to large environmental changes such as being affected, the illuminance in the entire image will be different, for example, when the input image is captured and when the background image is captured due to the gradual change of the environment over time. There was a possibility of outputting incorrect detection results.

  For this reason, the present invention can detect an intruder or the like in an entry-prohibited area such as an important facility or road while preventing an erroneous recognition due to an environmental change with respect to the entire input image using the image of the surveillance camera. It is an object of the present invention to provide an environment-adaptive intruder detection device using possible image processing.

An environment-adaptive intruder detection device based on image processing according to a first invention for solving the above-described problem is a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion, and In an environment-adaptive intruder detection device by image processing comprising an image processing unit for analyzing the input image transmitted from a monitoring camera, the image processing unit receives the input image from the storage means and the monitoring camera. An image input means for inputting and storing in the storage means, a plurality of change inspection small areas are set by dividing the input image, and a corresponding area of a background image previously obtained for each of the change inspection small areas An area where the difference from the background image is larger than a preset value is detected as a small change area by comparison, and an area different from the small change area of the change inspection small area is detected as a background sign. In the image change detection means for detecting as a small area, the sequential background update means for re-sequentially updating the background image data based on the preset condition for each background candidate small area, and the image change detection means comprising a intruder detection means for detecting an intruder based on the output the changed sub-regions, and a result output means for outputting a result detected in said intruder detection means, said sequentially background updating means, respectively A clock counter having a function of monitoring the time corresponding to the change inspection small area, and resetting the time counter when starting the image processing unit, and then monitoring the elapsed time and detecting it as the background candidate small area The background image corresponding to the corresponding background candidate small region when the time counter corresponding to the selected small region reaches a predetermined value set in advance It resets the time counter corresponding to the background candidate small area extracted the background image data with take heart background image data in the area, as the predetermined value, respectively is set arbitrarily for each said background candidate small areas A characteristic value is used.

An environment-adaptive intruder detection device based on image processing according to a second invention for solving the above-described problem is a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion, and In an environment-adaptive intruder detection device by image processing comprising an image processing unit for analyzing the input image transmitted from a monitoring camera, the image processing unit receives the input image from the storage means and the monitoring camera. An image input means for inputting and storing in the storage means, a plurality of change inspection small areas are set by dividing the input image, and a corresponding area of a background image previously obtained for each of the change inspection small areas An area where the difference from the background image is larger than a preset value is detected as a small change area by comparison, and an area different from the small change area of the change inspection small area is detected as a background sign. In the image change detection means for detecting as a small area, the sequential background update means for re-sequentially updating the background image data based on the preset condition for each background candidate small area, and the image change detection means An intruder detection means for detecting an intruder based on the output small change area; and a result output means for outputting a result detected by the intruder detection means; and the sequential background update means, The background image data of the region of the background image corresponding to the background candidate small region is updated based on the order set in advance for the change inspection small region.

An environment-adaptive intruder detection device based on image processing according to a third aspect of the present invention for solving the above-described problem is a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion; and In an environment-adaptive intruder detection device by image processing comprising an image processing unit for analyzing the input image transmitted from a monitoring camera, the image processing unit receives the input image from the storage means and the monitoring camera. An image input means for inputting and storing in the storage means, a plurality of change inspection small areas are set by dividing the input image, and a corresponding area of a background image previously obtained for each of the change inspection small areas An area where the difference from the background image is larger than a preset value is detected as a small change area by comparison, and an area different from the small change area of the change inspection small area is detected as a background sign. In the image change detection means for detecting as a small area, the sequential background update means for re-sequentially updating the background image data based on the preset condition for each background candidate small area, and the image change detection means An intruder detection means for detecting an intruder based on the outputted small change area; and a result output means for outputting a result detected by the intruder detection means. A plurality of the small areas adjacent to each other are set as one block in the input image, a ranking is set in advance for each of the block and the background candidate small areas in the block, The background image data of the area of the background image corresponding to the background candidate small area is updated based on the ranking.

  According to the present invention, it is possible to always compare a new background image data with an input image even when a place where a large environmental change with time elapses such as an outdoor environment is set as a monitoring target place, and adapts to the environmental change. A correct detection result can be obtained.

  In addition, since a correct detection result can be obtained by adapting to environmental changes over time, an outdoor monitoring system capable of continuous operation corresponding to daily weather changes can be constructed.

  Furthermore, since the monitoring device automatically responds to the environment in response to environmental changes, it is not necessary for the operator to perform initial system settings each time the weather changes, which can greatly improve convenience. .

  Embodiments of the present invention will be described in detail in the following examples.

  A first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a block diagram showing a configuration of an environment-adaptive intruder detection device using image processing according to the present embodiment, FIG. 2A is an explanatory diagram showing an example of dividing an input image into change inspection subregions, and FIG. FIG. 2C is an explanatory diagram showing an example of the change detection area, and FIG. 3 is an environment adaptive intruder by image processing according to the present embodiment. It is a flowchart which shows the flow of the process in a detection apparatus. The environment-adaptive intruder detection device by image processing according to the present embodiment detects an intruder into a prohibited area by performing image analysis on a surveillance camera image installed in a prohibited area such as a facility or a road. .

  As shown in FIG. 1, in this embodiment, the environment-adaptive intruder detection device using image processing includes a monitoring camera 1 and an image processing unit 2. The surveillance camera 1 is installed at a position overlooking the restricted entry area, images the prohibited entry area and its peripheral part, and transmits the image data to the image processing part 2. The image processing unit 2 performs an analysis process on the image captured by the monitoring camera 1 and outputs a detection result and a processing state. At least an image input unit 21 as an image input unit, an image change detection Image change detection unit 22 as means, sequential background update unit 23 as sequential background update means, intruder detection unit 24 as intrusion detection means, and result output unit 25 as result output means, and storage means As a processing memory 26.

  The image input unit 21 is means for inputting data of an image (hereinafter referred to as an input image) captured by the monitoring camera 1 from the monitoring camera 1 and storing it in the processing memory 26.

  Further, the image change detection unit 22 takes out the input image and the previously acquired background image data as image data from the processing memory 26 and compares them to detect a difference (change to the background image in the input image). The detected result is stored in the processing memory 26 as small area data. More specifically, as shown in FIG. 2A, the image change detection unit 22 divides the input image 3 taken out from the processing memory 26 and divides the input image 3 into a plurality of (108 in FIG. 31 (referred to as a change inspection subregion) is set, and for each change inspection subregion 31, correlation calculation between the change inspection subregion 31 and the region of the background image corresponding to the change inspection subregion 31 is performed. Then, from the change inspection small area 31 set in the input image 3 based on the temporal change of the correlation value, the change small area (the occurrence of a change due to the presence of the intruder 4 or the environmental change) ( For example, a hatched portion in FIG. 2B) 31a and a background candidate small region (a portion other than the small change region 31a) 31b with little change from the background image are detected, and the small change region 31a and the small background candidate are detected. Information on the area 31b is stored in the processing memory 26 as small area data.

  The sequential background update unit 23 is a unit that sequentially updates the background image based on preset conditions. Specifically, the sequential background update unit 23 corresponds to each change inspection small region 31 set in the image change detection unit 22 and the time during which the background image region corresponding to each change inspection small region 31 is not updated. A time counter (not shown) for monitoring is provided. Then, the data of the input image and the small area data are extracted from the processing memory 26 as the image processing unit 2 is activated, and after resetting the time counter, the time counter corresponding to the small area detected as the background candidate small area 31b. When the elapsed time monitored by the time counter reaches a predetermined value set in advance, the background image data corresponding to the background candidate small area 31b is re-acquired. When the background image data is re-acquired, the time counter corresponding to the corresponding background candidate small area 31b is reset at the same time. The acquired background image data is stored in the processing memory 26.

  The intruder detection unit 24 is a means for extracting small area data, which is information of changes detected by the image change detection unit 22, from the processing memory 26, and detecting an intruder based on the small area data. More specifically, the intruder detection unit 24 takes out the small area data stored in the processing memory 26 and combines adjacent change small areas 31a in the input image 3 as shown in FIG. Determining the change detection area 31c, and determining that the intruder exists in the change detection area 31c when the geometric feature amount of the change detection area 31c is within a feature amount range corresponding to a preset intruder. Thus, the intruder is detected and the position in the image is specified. The identified result is stored in the processing memory 26 as intrusion data.

  The result output unit 25 is a means for outputting the result detected by the intruder detection unit 24 (intruder data in this embodiment) and the like. The processing memory 26 stores various data output from the processing units 21 to 24, a background image, and the like. The background image is an image in which the same area as that captured by the monitoring camera 1 is captured in advance.

  Hereinafter, the flow of processing by the image processing unit 2 of the present embodiment shown in FIG. 1 will be described in detail with reference to FIG. As shown in FIG. 3, in the present embodiment, the image processing unit 2 first resets the time counter, and inputs image data of a prohibited entry imaged by the monitoring camera 1 by the image input unit 21 (step S <b> 101). ).

  Subsequently, the image change detection unit 22 divides the input image into a plurality of change inspection subregions 31 as shown in FIG. 2A, and the image change detection unit 22 inputs the background image as described above. A process of detecting a small change area 31a and a background candidate small area 31b in which an image has changed is performed (step S102).

  Thereafter, the sequential background update unit 23 monitors the time during which the background image region corresponding to the background candidate small region 31b detected by the image change detection unit 22 is not updated, and sets the time monitored by the time counter in advance. For the background image region corresponding to the background candidate small region 31b determined to have reached the predetermined value set in advance as a result of the determination, for example, at that time The background candidate small area 31b of the input image input in is fetched as background image data. On the other hand, as a result of the determination, the current background image is maintained for the region of the background image corresponding to the background candidate small region 31b determined that the time counter has not reached the predetermined value set in advance (step S103).

  Subsequently, the intruder detection unit 24 performs a process of detecting an intruder in the input image 3 (step S104). Finally, the result output unit 25 performs a process of outputting the presence / absence of the intruder detection (step S105). . That is, based on the entire illuminance detection state and the intruder data extracted from the processing memory 26, the presence / absence of the entire illuminance change detection, the presence / absence of the intruder detection, and the intruder data are output. For example, when the worker detects a change in overall illuminance, the worker can take necessary measures such as manually recovering the intruder detection process.

  According to the environment-adaptive intruder detection apparatus using image processing according to the above-described embodiment, the time counter is provided for the background candidate small area 31b by the time counter provided in each change inspection small area 31, and the time counter Since the background image data is re-acquired for the background image region corresponding to the background candidate small region 31b every time the predetermined value reaches a predetermined value set in advance, the small region gradually increases with time. The background image data can be updated, and the input image can always be compared with the new background image data even when the location subject to monitoring is a place where the environmental change with time is large, such as the outdoor environment. Therefore, a more accurate detection result adapted to the environmental change can be obtained.

(Method with added function to set offset time at startup)
A second embodiment of the present invention will be described. The present embodiment is an example in which the processing in the background update unit 23 is different from the environment-adaptive intruder detection device by image processing of the first embodiment. Specifically, in this embodiment, instead of resetting the time counter when the image processing unit 2 is activated in the sequential background update unit 23, each corresponding to each change inspection subregion 31 shown in FIG. A random offset time is set as a predetermined time until the background image is updated with respect to the time counter. Other configurations are generally the same as those described in the first embodiment, and redundant description is omitted.

  According to the environment-adaptive intruder detection device based on image processing according to the present embodiment, in addition to the effects of the device described in the first embodiment, a random offset time is set for the time counter of each small region 31 at the time of activation. Therefore, the value of each time counter varies, and it is possible to prevent background update from being performed simultaneously in the background image region corresponding to each background candidate small region 31b, and by updating the background image data It is possible to prevent the intruder detection process from being interrupted.

(Method to specify background update small area sequentially)
A third embodiment of the present invention will be described with reference to FIG. FIG. 4 is an explanatory diagram showing an example of designating a small region for background update in the environment adaptive intruder detection device based on image processing according to the present embodiment.

  The present embodiment is an example in which the processing in the background update unit 23 is different from the environment-adaptive intruder detection device by image processing of the first embodiment. Specifically, in this embodiment, the background image data is sequentially updated in the plurality of change inspection subregions 31 set in the input image 3 by the image change detector 22 as shown in FIG. (In Fig. 4, first specify sequentially from the left end of the first step toward the right end, and when the first step is completed, specify the order from the left end to the right end in the order of the second step and the third step, respectively. The background image data is updated by continuously or intermittently taking back the background image for the background image region corresponding to the background candidate small region 31b in the designated order.

  According to the environment-adaptive intruder detection apparatus based on image processing according to the present embodiment, every time a small area is designated as a small area for background acquisition, the background image data of that area is re-acquired. Because the background image data for each small area is updated, it is always possible to compare the input image with the new background image data even when the place where the environmental change is large over time, such as the outdoor environment, is the monitoring target place. It is possible to obtain correct detection results adapted to environmental changes. In addition to this, since the small areas for actual background update are sequentially specified, it is possible to reliably avoid the background update of each small area at the same time, and the intruder detection process is interrupted by updating the background. Can be prevented.

(Method to specify background update small area sequentially by block)
A fourth embodiment of the present invention will be described with reference to FIG. FIG. 5 is an explanatory diagram illustrating an example of designating a small region for background update in the environment adaptive intruder detection device based on image processing according to the present embodiment.

  The present embodiment is an example in which the processing in the sequential background update unit 23 is different from the first embodiment. Specifically, in this embodiment, the sequential background update unit 23 uses a plurality of (9 in FIG. 5) small regions 31 as one region (hereinafter referred to as a block) 32 as shown in FIG. A plurality (9 in FIG. 5) of the blocks 32 are set, and the order of updating the background is designated for the block 32 and the background candidate small area 31b in one block 32, respectively.

  Describing in detail using the input image 3 shown in FIG. 5 as an example, the background update order is first set to the left end of the first stage (uppermost stage) for the background candidate small area 31b in one block 32. From the left end to the right end in the same order in the second and third stages. For the block 32, first, the first row (leftmost column) is designated sequentially from the upper end to the lower end, and when the first row is completed, the second row and the third row are sequentially followed from the upper end to the lower end. The direction is specified sequentially.

  That is, first, the background image data is re-acquired in the order described above for the background image areas corresponding to all the background candidate small areas 31b in the block 32 at the upper left of the input image 3 shown in FIG. After that, the background image data in the order of the background image data corresponding to all the background candidate small areas 31b in the block 32 (assumed to be a block 32-2) adjacent to the lower side of the block 32-1. The background image is sequentially updated in such a manner that the image is taken again.

  According to the environment-adaptive intruder detection device using image processing according to the present embodiment described above, in addition to the effects of the device shown in the third embodiment, the block 32 for the region of the background image corresponding to the background candidate small region 31b. Since the order of re-acquisition of the background image is specified every time, the region where the data is updated in the background image is always prevented from being continuous, and even when the detection target happens to exist in the small region where the background is updated, The intruder detection can be performed with high accuracy without affecting the intruder detection processing by the background update.

  The present invention relates to an apparatus for detecting an intruder into a facility or road entry prohibition place, and in particular, by installing a surveillance camera for monitoring the entry prohibition place and processing the captured image, This can be applied to an intruder detection device that detects a person or the like who enters a place where no entry is permitted.

It is a block diagram which shows schematic structure of the environment adaptive type intruder detection apparatus by the image processing which concerns on Example 1 of this invention. 2A is an explanatory diagram showing an example of dividing an input image into change inspection subregions, FIG. 2B is an explanatory diagram showing examples of change subregions and background candidate subregions, and FIG. It is explanatory drawing which shows the example of a detection area. It is a flowchart which shows the flow of the intruder detection process in the environment adaptive type intruder detection apparatus by the image processing which concerns on Example 1 of this invention. It is explanatory drawing which shows the example of the designation | designated order of the background update in the environment adaptive type intruder detection apparatus by the image processing which concerns on Example 3 of this invention. It is explanatory drawing which shows the example of the designation | designated order of the background update in the environment adaptive type intruder detection apparatus by the image processing which concerns on Example 4 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Surveillance camera 2 Image processing part 3 Input image 21 Image input part 22 Image change detection part 23 Sequential background update part 24 Intruder detection part 25 Result output part 26 Processing memory 31 Change inspection small area 31a Change small area 31b Background candidate small area 31c Change detection area 32 blocks

Claims (3)

Environment adaptation by image processing comprising: a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion; and an image processing section for analyzing the input image transmitted from the surveillance camera In type intruder detection device,
The image processing unit is configured to store a storage unit, an image input unit that inputs the input image from the monitoring camera and stores the input image in the storage unit, and sets a plurality of change inspection subregions by dividing the input image, The change inspection small region is compared with the corresponding region of the background image obtained in advance to detect a region whose difference from the background image is larger than a preset value as a change small region, and the change inspection small region Image change detecting means for detecting an area different from the small change area as a background candidate small area, and sequentially re-extracting the background image based on a preset condition for each background candidate small area and sequentially updating the background image data Background update means, intruder detection means for detecting an intruder based on the small change area output by the image change detection means, and detection by the intruder detection means. And a result output means for outputting a result of the,
The sequential background update means includes a clock counter having a function of monitoring time corresponding to each of the change inspection small areas, and monitors the elapsed time after resetting the time counter when the image processing unit is activated. When the time counter corresponding to the small area detected as the background candidate small area reaches a predetermined value set in advance, the background of the area of the background image corresponding to the background candidate small area Resetting the image data and resetting the time counter corresponding to the background candidate small area from which the background image data was extracted, and as the predetermined value, a value arbitrarily set for each of the background candidate small areas environment-adaptive intruder detecting apparatus according to images processed you characterized by using. Environment adaptation by image processing comprising: a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion; and an image processing section for analyzing the input image transmitted from the surveillance camera In type intruder detection device,
The image processing unit is configured to store a storage unit, an image input unit that inputs the input image from the monitoring camera and stores the input image in the storage unit, and sets a plurality of change inspection subregions by dividing the input image, The change inspection small region is compared with the corresponding region of the background image obtained in advance to detect a region whose difference from the background image is larger than a preset value as a change small region, and the change inspection small region Image change detecting means for detecting an area different from the small change area as a background candidate small area, and sequentially re-extracting the background image based on a preset condition for each background candidate small area and sequentially updating the background image data Background update means, intruder detection means for detecting an intruder based on the small change area output by the image change detection means, and detection by the intruder detection means. And a result output means for outputting a result of the,
The successive background update means, said change examination small area images you and updates the background image data in the area of the background image corresponding to the background candidate small area on the basis of the order to be set in advance for the Environment-adaptive intruder detection device by processing. Environment adaptation by image processing comprising: a surveillance camera configured to transmit an input image obtained by imaging an intrusion prohibited area and its peripheral portion; and an image processing section for analyzing the input image transmitted from the surveillance camera In type intruder detection device,
The image processing unit is configured to store a storage unit, an image input unit that inputs the input image from the monitoring camera and stores the input image in the storage unit, and sets a plurality of change inspection subregions by dividing the input image, The change inspection small region is compared with the corresponding region of the background image obtained in advance to detect a region whose difference from the background image is larger than a preset value as a change small region, and the change inspection small region Image change detecting means for detecting an area different from the small change area as a background candidate small area, and sequentially re-extracting the background image based on a preset condition for each background candidate small area and sequentially updating the background image data Background update means, intruder detection means for detecting an intruder based on the small change area output by the image change detection means, and detection by the intruder detection means. And a result output means for outputting a result of the,
The sequential background updating unit sets a plurality of blocks in the input image with a plurality of small regions adjacent to each other as one block, and for each of the block and the background candidate small regions in the block previously setting the rank, the environment adaptive intruder detecting apparatus according to images processed you and updates the background image data in the area of the background image corresponding to the background candidate small area on the basis of the order Te.

Images