JP2011053005A - Monitoring system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system capable of measuring a distance to an intruder, without using an exclusive device such as a distance-measuring sensor. <P>SOLUTION: A monitoring system 3 includes a monitoring device 1 and a surveillance camera 2. From an image photographed by the surveillance camera 2, the monitoring device 1 creates a distance image. When the monitoring device becomes in a monitoring state, the monitoring device 1 detects an intruder from the image photographed by the surveillance camera 2. When the monitoring device detects the intruder, the monitoring device 1 subsequently identifies a grounding position on the image of the intruder. When the monitoring device 1 identifies the grounding position, the distance to the grounding position is identified by lapping the grounding position on the distance image created previously. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technique for measuring a distance to an intruder. Specifically, the present invention relates to a technique for measuring a distance to an intruder using an image captured by a surveillance camera.

  Surveillance cameras are installed in facilities such as buildings and houses. An intruder can be detected by analyzing the image captured by the surveillance camera. When an intruder is detected, countermeasures such as sounding an alarm are taken. Alternatively, it is possible to grasp the situation in the facility by recording an image captured by the surveillance camera and later viewing the recorded captured image.

  There is a technique for measuring the distance to a person. For example, there are a technique for measuring a distance to a person using a laser sensor and a technique for measuring a distance to a person using a stereo image. Since a dedicated device is required for these, the system cost is relatively high.

  In the following Patent Document 1 and Patent Document 2, a technique for measuring a distance to a person using a captured image is disclosed.

  In Patent Document 1, a distance to a person is specified by comparing an image of a reference object whose height and position are known with a person image. When comparing the image of the reference object and the person image, an evaluation independent of the height of the person is performed by using the distance from the vanishing point to the lower end of the object.

  In Patent Document 2, a conversion table that associates the position of the head image on the image with the coordinates of the floor portion is used. By specifying the position of the head from the captured image, the position of the person is specified using the conversion table.

JP 2008-286638 A Japanese Patent No. 3637360

  An object of the present invention is to provide a technique for measuring a distance to an intruder using a captured image without separately providing a dedicated device for distance measurement such as a laser sensor.

  According to the first aspect of the present invention, an imaging device in which an imaging range is set so as to capture a predetermined region, and coordinates on the image of the predetermined region captured by the imaging device from the imaging device in real space. A storage device that stores a distance image corresponding to a distance, and an intruder detection unit that acquires a monitoring image of the predetermined region by the imaging device and detects an intruder from the monitoring image while monitoring the predetermined region A grounding position identifying unit that identifies the grounding position of the intruder in the monitoring image; and a distance identifying unit that identifies a distance from the distance image and the grounding position to the intruder. To do.

  According to a second aspect of the present invention, in the monitoring system according to the first aspect, the intruder detection unit compares the monitoring image with a reference image captured without an intruding object in the predetermined area. And an extraction unit that extracts an area where the change in luminance value exceeds the first threshold as a moving body area, and when the height of the moving body area exceeds a second threshold, the moving body area is And an intruder determination unit that determines the area.

  According to a third aspect of the present invention, in the monitoring system according to the first aspect, the intruder detection unit is configured such that the difference between the monitoring image and a reference image captured without an intruding object in the predetermined area. A difference generating unit that generates an image, and a matching unit that performs matching between the difference image and a specific pattern image that represents a feature of a specific place of a person, and determines the area of the intruder included in the monitoring image It is characterized by including these.

  According to a fourth aspect of the present invention, in the monitoring system according to the second aspect, the grounding position specifying unit specifies the lowest point of the moving body region as the grounding position.

  According to a fifth aspect of the present invention, in the monitoring system according to the third aspect, when the ground contact position specifying unit succeeds in matching the difference image with a foot pattern image representing a feature of a person's foot, the foot The lowest point of the area matched with the pattern image is specified as the ground contact position.

  According to a sixth aspect of the present invention, in the monitoring system according to the second aspect, when the lowest point of the moving body region coincides with the lowest region of the monitoring image, the grounding position specifying unit , It is determined that the camera is approaching within a predetermined distance from the imaging device.

  According to a seventh aspect of the present invention, in the monitoring system according to the second aspect, the intruder detection unit includes a size storage unit that stores a size of the detected intruder, and the grounding position specifying unit is a current If the size of the intruder included in the frame is smaller than the intruder size included in the previous frame one frame before by a predetermined percentage or more, the position of the head of the intruder included in the current frame is determined. As a reference, the ground contact position is specified by considering the size of the intruder included in the past frame.

  The invention according to claim 8 is the monitoring system according to claim 2, wherein the intruder detection unit includes a size storage unit that stores a size of the detected intruder, and the grounding position specifying unit is a current When the size of the intruder included in the frame is larger than the size of the intruder included in the previous frame one frame before, a size of the intruder included in the past frame is considered. And the ground contact position is specified based on the current frame.

  According to a ninth aspect of the present invention, in the monitoring system according to the third aspect, in the lower end region of the difference image, the grounding position specifying unit is configured to perform a difference between the difference image and an upper body pattern image representing a characteristic of the upper body of the person. When the matching is successful, it is determined that the intruder is approaching within a predetermined distance from the imaging device.

  According to a tenth aspect of the present invention, in the monitoring system according to the third aspect, when the ground contact position specifying unit fails to match the difference image with a foot pattern image representing a feature of a person's foot, matching is performed. The ground contact position is specified by applying a whole person pattern image representing the characteristics of the whole person to another place of the person who has succeeded.

  The invention according to claim 11 is the monitoring system according to claim 3, further comprising a ratio storage unit that stores ratio information indicating a ratio of an average figure of the person, When matching between the difference image and the foot pattern image representing the characteristics of the person's foot could not be performed, by applying the ratio indicated by the ratio information to the size of the other place of the person who succeeded in matching, The ground contact position is specified.

  The invention according to claim 12 is the monitoring system according to claim 3, further comprising a ratio storage unit that stores ratio information indicating a ratio between the average face size of the person and the size of the whole body, The contact position specifying unit performs face recognition processing on the monitoring image, and cannot perform matching between the face extraction unit that acquires the size of the face, the difference image, and a foot pattern image that represents a feature of a person's foot In this case, the ground contact position is specified by applying the ratio indicated by the ratio information to the size of the face extracted by the face extraction unit.

  By using the monitoring system of the present invention, the distance to the intruder can be measured using the captured image without using a distance measuring sensor or the like.

1 is an overall view of a monitoring system. It is a block diagram of a monitoring apparatus. It is a figure which shows the mark object installed in the monitoring area. It is a figure which shows a distance image. It is a figure which shows the mark line arrange | positioned in the monitoring area. It is the elements on larger scale of the surveillance picture containing a person. It is the elements on larger scale of the luminance image containing a person. It is a figure which shows the lowest point of the person extracted from the brightness | luminance image. It is a figure which shows the lowest point of the person extracted from the brightness | luminance image. It is a figure which shows the monitoring image in connection with before and behind a person's movement. It is a figure which shows the brightness | luminance image in connection with before and behind a person's movement. It is a figure which shows the grounding position shown on the distance image. It is a figure which shows a contact position and a monitoring area boundary line. It is a figure which shows a reference | standard image. It is a figure which shows the monitoring image containing a person. It is a figure which shows the difference image of a reference | standard image and a monitoring image. It is a figure which shows the matching method of a difference image and a person pattern image. It is a figure which shows the matching method of a difference image and a foot pattern image. It is a figure which shows the monitoring image in which a person is located in the lower end of an imaging area. It is a figure which shows the monitoring image of the state where the person was hidden behind the obstacle. It is a figure which shows the state which applied the image of the lower body to the image of the upper body. It is a figure which shows the method of estimating the length of the whole body from the length of an upper body. It is a figure which shows the method of estimating the length of the whole body from the length of a face.

{First embodiment}
Hereinafter, a monitoring system 3 according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is an overall view of the monitoring system 3. The monitoring system 3 is set to monitor the site of the building 5. The monitoring system 3 includes a monitoring device 1 and a monitoring camera 2 installed in a building 5.

  As shown in FIG. 2, the monitoring device 1 includes a control unit 11 including a CPU, a memory, and the like, and a storage device 12. A monitoring camera 2 is connected to the control unit 11. The surveillance camera 2 includes a CCD image sensor, for example, and can image a peripheral region of the building 5. The monitoring camera 2 has a fixed height, direction, and the like so that the monitoring area 6 shown in FIG. 1 can be imaged. In this way, the surveillance camera 2 captures an image within the fixed imaging range.

Hereinafter, regarding the operation and configuration of the monitoring system 3 that specifies the distance from the monitoring camera 2 to the intruder,
<1. Creating a distance image>
<2. Intruder detection>
<3. Identifying the grounding location>
<4. Identification of distance>
Will be described in the order.

<1. Creating a distance image>
The distance image creation process will be described. Before actually operating the monitoring system 3 of the present embodiment, a distance image is created in advance.

  In order to create a distance image, a plurality of landmark objects 71, 72, and 73 for distance measurement are installed in the monitoring area 6 as shown in FIG. The mark object 71 is installed at a location 15 m away from the installation location of the surveillance camera 2. The mark object 72 is installed at a distance of 10 m from the installation place of the monitoring camera 2. The mark object 73 is installed at a distance of 5 m from the installation place of the monitoring camera 2. The landmark objects 71, 72, and 73 are intended to measure the distance from the monitoring camera 2 by the height of the ground (vertical position in the image), so it is better to select an object that is as low as possible.

  After the landmark objects 71, 72 and 73 are arranged in the monitoring area 6, an image 81 is taken by the monitoring camera 2. As shown in FIG. 4, the captured image 81 includes landmark objects 71, 72, and 73. On this image 81, a line 811 indicating a point 15 m away from the surveillance camera 2 is drawn by connecting a plurality of landmark objects 71. Similarly, by connecting between the plurality of landmark objects 72 and between the landmark objects 73, lines 811 indicating points of a distance of 10 m and 5 m from the monitoring camera 2 are drawn.

  The distance image data 121 is generated using the image 81. The distance image data 121 has the same number of pixels as the image 81. Each pixel of the distance image data 121 is associated with a distance from the monitoring camera 2. In the distance image data 121, the distance 15m is associated with the pixels on the line 811 connecting the landmark objects 71. In the distance image data 121, distances 10m and 5m are associated with pixels on the line 811 connecting the mark objects 72 and the mark objects 73, respectively.

  Furthermore, as shown in FIG. 4, a line 811 corresponding to a distance of 12.5 m is drawn between the line 811 connecting the mark objects 71 and the line 811 connecting the mark objects 72 by estimation. Similarly, a line 811 such as a distance of 7.5 m and a distance of 17.5 m is also drawn by estimation. In the distance image data 121, distances are associated with the pixels on the line 811 created based on the landmark objects 71, 72, and 73 and the estimated pixels on the line 811.

  It is possible to freely set the intervals at which the mark objects are installed and the extent to which the line 811 is estimated in detail.

  Although the case where the mark objects 71, 72, and 73 are installed for distance measurement has been described as an example, a distance measurement line 75 may be drawn in the monitoring area 6 as shown in FIG. For example, a line 75 such as a white line may be drawn in the monitoring area 6. A mark point 76 is added to the line 75 at a distance of 5 m, 10 m, 15 m, or the like from the surveillance camera 2. In the distance image data 121, the distance is associated with the pixels on the line connecting the landmark points 76. Further, the estimated line is complemented between the lines connecting the mark points 76, and the distance is also associated with the pixels on the estimated line.

  The distance image data 121 is created by performing the above operations. The distance image data 121 is created by the monitoring program 120 executing based on the image data captured by the monitoring camera 2. The interval between the lines 811 can be appropriately set by the user from the input device.

  When the creation of the distance image data 121 is completed, the landmark objects 71, 72, and 73 may be removed. Alternatively, the line 75 and the mark line 76 may be deleted.

<2. Intruder detection>
After storing the distance image data 121 in the storage device 12, the monitoring system 3 shifts to the monitoring state. When entering the monitoring state, the monitoring camera 2 captures an image in the monitoring area 6 at a rate of 2 to 5 frames / second, for example. The captured image is stored in the storage device 12 as captured image data 122.

  The contents of the intruder detection process executed by the monitoring program 120 will be described below.

  FIG. 6 shows partially enlarged views 82N and 82 (N + 1) of the captured image data 122. FIG. The partial enlarged view 82N is an enlarged view of a person part in the captured image data 122 captured in the Nth frame. Partial enlarged view 82 (N + 1) is an enlarged view of a person portion in captured image data 122 captured in the (N + 1) th frame. It can be seen that the person is moving from the Nth frame to the (N + 1) th frame.

  FIG. 7 shows a luminance image. The luminance image 92N is a partially enlarged view of the luminance image of the Nth frame. That is, it is a luminance image corresponding to the partially enlarged view 82N of FIG. The luminance image 92 (N + 1) is a partially enlarged view of the luminance image of the (N + 1) th frame. That is, the luminance image corresponds to the partial enlarged view 82 (N + 1) in FIG.

  The numerical values shown in the luminance images 92N and 92 (N + 1) are luminance values. However, the luminance values shown in the figure are not actual luminance values. In order to simplify the explanation, the luminance values shown in the figure are expressed by numerical values as small as one digit, and the larger the numerical value, the higher the luminance. In the luminance image, luminance values may be associated with all pixels, but in the examples illustrated in FIGS. 6 and 7, luminance values are associated with blocks composed of a plurality of pixels. For example, as the luminance value of each block, an average value of luminance values of all pixels included in the block may be used.

  The monitoring program 120 compares the luminance image 92N and the luminance image 92 (N + 1), and identifies an area where the luminance change exceeds a predetermined threshold as the moving body area 921. In FIG. 7, the moving body region 921 is indicated by hatching. Specifically, an area where the luminance value has increased by a predetermined threshold or more as the Nth frame changes to the (N + 1) th frame is specified as the moving body area 921.

  Subsequently, the monitoring program 120 determines whether or not the height of the moving object region 921 exceeds a predetermined threshold value. The monitoring program 120 determines that the moving body area 921 is an intruder detection area when the height of the moving body area 921 exceeds a predetermined threshold.

  If the height of the moving object region 921 is below a predetermined threshold, it is determined that there is no intruder, and the process moves to the next frame. The case where the height of the moving body region 921 falls below a predetermined threshold value may be, for example, a case where a dog or cat has entered the monitoring area 6. In such a case, it is possible to prevent erroneous detection that there is an intruder.

<3. Identifying the grounding location>
When the presence of the intruder is confirmed from the mobile object area, the monitoring program 120 next specifies the grounding position of the intruder.

  FIG. 8 shows a luminance image 93A. A moving body region 931 is detected in the luminance image 93A. Since the height of the mobile body region 931 exceeds a predetermined threshold, it is determined that the mobile object region 931 is an intruder. The monitoring program 120 determines that the lowest position 932 of the moving body region 931 is the ground contact position of the intruder.

  In the example shown in FIG. 8, the entire moving body region 931 is within the imaging range. That is, the entire body of the intruder is within the imaging range of the surveillance camera 2. On the other hand, in the example shown in FIG. 9, the moving body region 931 is located at the lower end in the imaging range. That is, only the upper body of the intruder is within the imaging range.

  In such a case, it is inappropriate to determine the lowest position 934 of the moving body region 933 as the ground contact position of the intruder. In such a case, the monitoring program 120 does not determine the lowest position 934 as the ground contact position. The monitoring program 120 determines that an intruder is approaching an area where the distance from the monitoring camera 2 is a predetermined value or less. For example, it is determined that the intruder is within a range of 5 m from the surveillance camera 2.

  Another case will be described. 10 and 11 are diagrams illustrating a case where an intruder is hidden behind an obstacle. FIG. 10 is a diagram showing a frame image 83C and a frame image 83D. The frame image 83C is an Nth frame image, and the frame image 83D is an (N + 1) th frame image.

  FIG. 11 is a diagram showing a luminance image 93C and a luminance image 93D. The luminance image 93C is a luminance image corresponding to the frame image 83C. That is, the luminance image of the Nth frame. The luminance image 93D is a luminance image corresponding to the frame image 83D. That is, it is a luminance image of the (N + 1) th frame.

  As shown in FIG. 10, the whole body of the intruder 831 is captured by the monitoring camera 2 in the frame image 83C. Correspondingly, as shown in FIG. 11, the moving body region 931 is also within the imaging range in the luminance image 93 </ b> C.

  As shown in FIG. 10, in the frame image 83D, the intruder 831 is hidden behind the obstacle, and only the upper body is imaged by the surveillance camera 2. Correspondingly, as shown in FIG. 11, the size of the moving object region 931 in the luminance image 93D is smaller than the size of the moving object region 931 in the luminance image 93C.

  In the luminance image 93D shown in FIG. 11, it is inappropriate to set the lowest position as the ground contact position. As described above, when the size of the moving object region is reduced, the monitoring program 120 uses the size of the moving object region of the previous frame.

  As shown in FIG. 2, the monitoring device 1 stores size data 123 in the storage device 12. The size data 123 is the size of the moving object area. As the size of the mobile object region, the height of the mobile object region may be used. Each time the moving body area is detected, the storage device 12 stores size data 123 of the moving body area. The size data 123 is recorded so that the latest data is updated.

  The monitoring program 120 applies the size of the moving body area of the previous frame to the current frame when the size of the moving body becomes smaller than a predetermined ratio. Specifically, the highest position of the moving body area of the current frame is regarded as the head. Then, the area is estimated downward from the highest position by the size of the moving object area acquired in the previous frame, and the lowest position of the estimated area is set as the ground contact position. In this way, even when an intruder is hidden by an obstacle, the ground contact position can be estimated.

  The monitoring program 120 analyzes the moving body area of the current frame without applying the moving body area size of the previous frame when the size of the moving body becomes larger than a predetermined ratio. And the lowest position of a mobile body area | region is pinpointed as a grounding position by performing the process as shown in FIG.

<4. Identification of distance>
The monitoring program 120 specifies the ground contact position of the intruder by executing the above processing. When the contact position is specified, the contact position is applied to the distance image data 121, and the distance from the monitoring camera 2 at the position where the intruder exists is specified.

  FIG. 12 is a diagram illustrating a method of specifying the distance. The ground contact position 87 is applied on the distance image data 121, and the ground contact position of the intruder, that is, the distance from the monitoring camera 2 is specified. If the ground contact position 87 is on the line 811, the distance directly associated with the line 811 is the distance to the ground contact position.

  If the contact position 87 is not on the line 811, the distance is estimated from the positional relationship between the line 811 and the contact position 87 as shown in the figure. In the illustrated example, the distance 871 is a distance from the ground contact position 87 to 12.5 m, and the distance 872 is a distance from the ground contact position 87 to 10 m. The point that internally divides the distance of 10 m and the distance of 12.5 m from the ratio of the distance 871 and the distance 872 is considered as the ground contact position, and the distance is obtained. Alternatively, in order to simplify the processing, when the ground contact position 87 is not on the line 811, the closest distance on the line 811 may be applied.

  As shown in FIG. 13, it is assumed that a distance 13 m is a monitoring area boundary line 88. In this case, when an intruder enters near the distance 13 m, it is necessary to enter a warning state as an intruder exists. For example, when the ground contact position of the mobile object region is the ground contact position 87A in the figure, it is farther than the 13 m monitoring area boundary line 88, so that it does not transition to the alert state. When the grounding position of the moving body is the grounding position 87B in the figure, it is closer to the monitoring area boundary line 88 of 13 m, so that it shifts to a warning state.

  When entering the alert state, the monitoring system 1 performs various operations. For example, an alarm sound is sounded to threaten the intruder and notify that there are intruders around. Alternatively, the warning light may be turned on. Further, the frame rate of the captured image data 121 may be increased and stored.

{Second Embodiment}
Next, a second embodiment of the present invention will be described. In the second embodiment, an intruder detection method is different.

  In the second embodiment, <1. Creation of range image>, <4. Specifying the distance> is the same as in the first embodiment. Hereinafter, <2. Intruder detection> and <3. Next, the specification of the grounding position will be described.

<2. Intruder detection>
FIG. 14, FIG. 15 and FIG. 16 are diagrams showing an intruder detection method. FIG. 14 shows a reference image 83E obtained by imaging the imaging area 6. The reference image 83E is captured in advance before entering the monitoring state. The reference image 83E is captured in a state where there is no moving body such as a person in the imaging area 6. The reference image 83E is stored in the storage device 12 as reference image data 124 as shown in FIG.

  With the reference image 83E stored, the surveillance camera 2 starts imaging the imaging area 6. The captured image is stored in the storage device 12 as captured image data 122. FIG. 15 shows a monitoring image 83 </ b> F captured while the intruder 831 has entered the imaging area 6.

  The monitoring program 120 extracts a difference between the monitoring image 83F and the reference image 83E, and generates a difference image 83G. FIG. 16 shows a difference image 83G. In the difference image 83G, as shown in the figure, the background portion is removed and the difference area 831 is extracted. In the following description, the intruder 831 will be referred to as a difference area 831 as appropriate.

  After generating the difference image 83G, the monitoring program 120 next performs matching with the difference image 83G using the pattern image extracted from the pattern image data 125.

  The pattern image data 125 is stored in the storage device 12 as shown in FIG. The pattern image data 125 includes a pattern image indicating the characteristics of the person's whole body. The monitoring program 120 can appropriately extract a pattern image at an arbitrary place from the pattern image data 125 indicating the characteristics of the whole body and use it for matching. For example, an arbitrary place such as a head pattern indicating a head feature, an upper body pattern indicating an upper body feature, and a foot pattern indicating a foot feature can be extracted in an arbitrary range and used as a pattern image.

  FIG. 17 is a diagram illustrating a matching method. For example, the upper body pattern 101 and the body pattern 102 are matched with the difference region 831. If any pattern matches the difference area 831, the monitoring program 120 determines that the difference area 831 is the intruder 831. If it matches any one of the patterns, it may be determined that the matching is successful. For example, if the patterns at a plurality of places such as two places match, it may be determined that the matching is successful.

<3. Identifying the grounding location>
If the monitoring program 120 determines that the difference area 831 is the intruder 831, the monitoring program 120 next specifies the ground contact position. FIG. 18 is a diagram illustrating a method for specifying the ground contact position. The monitoring program 120 performs matching between the difference area 831 and the foot pattern 103. When matching with the foot pattern 103 is possible, the lowest position of the matched foot can be specified as the ground contact position.

  In the example illustrated in FIG. 18, the entire body of the intruder 831 is within the imaging region. On the other hand, in the example shown in FIG. 19, a part of the intruder 831 is within the imaging area. That is, the difference area 831 is located at the lower end of the imaging range.

  As illustrated in FIG. 17, the monitoring program 120 extracts various feature patterns from the pattern image data 125 and performs matching with the difference area 831. In the case of FIG. 19, matching with the upper body pattern 101 is successful. When the matching is successful, the monitoring program 120 determines that the difference area 831 is the intruder 831.

  However, in the example shown in FIG. 19, the difference area 831 included in the difference image 83H is located at the lower end of the imaging range. In this case, the monitoring program 120 fails to match the difference area 831 and the foot pattern 103. The monitoring program 120 succeeds in matching with a feature pattern other than the foot pattern 103, but if the matching with the foot pattern 103 fails, the intruder 831 is positioned closer than the predetermined distance from the monitoring camera 2. Judge that. In other words, the monitoring program 120 determines that the intruder 831 is located closer than the predetermined distance from the monitoring camera 2 when the matching with the upper body pattern 101 is successful at the lower end position of the imaging region.

  FIG. 20 is a diagram showing another case. The monitoring image 83I is an image captured in a state where the intruder 831 is hidden behind an obstacle. In such a case, if the monitoring program 120 extracts a difference area and performs matching with the upper body pattern 101, matching is successful. However, the monitoring program 120 fails to match the difference area and the foot pattern 103.

  At this time, as shown in FIG. 21, the monitoring program 120 complements the foot portion by applying the foot pattern 103 to the difference region 831 of only the upper body included in the difference image 83J. Then, the lowest position of the foot of the foot pattern 103 used for this complementation is specified as the ground contact position.

  Here, it has been described that the foot pattern is complemented to the difference area 831 of only the upper body. However, since the pattern image data 125 substantially includes a pattern indicating the characteristics of the whole body, the difference area 831 of only the upper body is included. This is the same as the operation to superimpose the whole body pattern. By superimposing the whole body pattern on the difference area 831, that is, by superposing the whole body pattern while matching the upper body part, the lowest position of the foot can be specified as the ground contact position.

  FIG. 22 shows another method of specifying the contact position when matching with the foot pattern cannot be performed. This method is a method for identifying the ground contact position from the ratio of the part of the person included in the difference area 831 to the whole body. A difference area 831 of the upper body is extracted from the difference image 83K shown in FIG. And matching with the foot pattern has failed.

  The control unit 11 holds average body ratio data in advance. Therefore, the size of the whole body can be estimated from the size of the upper body based on the ratio data. Here, as the size, for example, the height of the region may be used.

  For example, if the height of the difference region 831 that is the upper body is L1, and the ratio of the height (length) of the upper body and the whole body is Ra, the whole body height L2 is expressed by L1 × Ra. The monitoring program 120 specifies the ground contact position using the estimated whole body height L2.

  FIG. 23 shows another method of specifying the contact position when matching with the foot pattern cannot be performed. This method is a method of specifying the ground contact position from the ratio of the face area and the whole body included in the difference area. For example, face recognition is performed from features such as eyes, nose, mouth, and ears, and the size of the face area is calculated.

  The control unit 11 holds ratio data between the face size and the whole body size in advance. Here, the size may be the height (length) of the face and the whole body. If the height of the face is L3 and the ratio between the height of the face and the whole body is Rb, the whole body height L4 is represented by L3 × Rb. The monitoring program 120 specifies the ground contact position from the height L4 of the whole body.

  As described above, according to the monitoring system 3 of the present embodiment, the distance to the intruder can be measured using the image captured by the monitoring camera 2.

  The storage device 12 of the monitoring device 1 shown in FIG. 2 stores all data so that both the first embodiment and the second embodiment can be processed. It is not limited to this. For example, if the monitoring system 3 only needs to execute the first embodiment, it is not necessary to store the reference image data 124 and the pattern image data 125.

1 Monitoring Device 2 Monitoring Camera 3 Monitoring System 6 Imaging Area

Claims (12)

An imaging device in which an imaging range is set to image a predetermined area;
A storage device that stores a distance image in which a distance from the imaging device in real space is associated with coordinates on the image of the predetermined region captured by the imaging device;
While monitoring the predetermined area, an intruder detection unit that acquires a monitoring image of the predetermined area by the imaging device and detects an intruder from the monitoring image;
A grounding position identifying unit that identifies the grounding position of the intruder in the monitoring image;
A distance specifying unit for specifying a distance from the distance image and the ground contact position to the intruder;
A monitoring system comprising: The monitoring system according to claim 1,
The intruder detection unit
An extraction unit that compares the monitoring image with a reference image captured in a state where there is no intruding object in the predetermined region, and extracts a region in which a change in luminance value exceeds a first threshold as a moving body region;
An intruder determination unit that determines the moving body area as the intruder area when the height of the moving body area exceeds a second threshold;
A monitoring system comprising: The monitoring system according to claim 1,
The intruder detection unit
A difference generating unit that generates a difference image between the monitoring image and a reference image captured in a state where there is no intruding object in the predetermined region;
A matching unit that performs matching between the difference image and a specific pattern image representing characteristics of a specific place of a person, and determines a region of the intruder included in the monitoring image;
A monitoring system comprising: The monitoring system according to claim 2,
The monitoring system characterized in that the grounding position specifying unit specifies the lowest point of the moving body region as a grounding position. The monitoring system according to claim 3,
The grounding position specifying unit is
When the difference image is successfully matched with a foot pattern image representing the characteristics of a person's foot,
A monitoring system, characterized in that the lowest point of an area matched with the foot pattern image is specified as a ground contact position. The monitoring system according to claim 2,
The ground contact position specifying unit determines that the intruder is approaching within a predetermined distance from the imaging device when the lowest point of the moving body region coincides with the lowest region of the monitoring image. A characteristic surveillance system. The monitoring system according to claim 2,
The intruder detection unit
A size storage unit for storing the size of the detected intruder,
Including
The grounding position specifying unit is included in the current frame when the size of the intruder included in the current frame is smaller than a predetermined ratio than the size of the intruder included in the previous frame of the previous frame. The monitoring system, wherein the ground contact position is specified by taking into account the size of the intruder included in the past frame with reference to the position of the head of the intruder. The monitoring system according to claim 2,
The intruder detection unit
A size storage unit for storing the size of the detected intruder,
Including
The grounding position specifying unit is included in the past frame when the size of the intruder included in the current frame is larger than a predetermined ratio than the size of the intruder included in the previous frame one frame before. The monitoring system, wherein the ground contact position is specified based on the current frame without considering the size of the intruder. The monitoring system according to claim 3,
In the lower end region of the difference image, the grounding position specifying unit, when the matching between the difference image and the upper body pattern image representing the upper body feature of the person is successful, the intruder is a predetermined distance from the imaging device A monitoring system characterized by being determined to be approaching within. The monitoring system according to claim 3,
If the difference image and the foot pattern image representing the feature of the person's foot could not be matched, the ground contact position specifying unit may display the whole person representing the feature of the whole person in another place of the person who succeeded in matching A monitoring system that identifies the ground contact position by applying a pattern image. The monitoring system according to claim 3, further comprising:
A ratio storage unit that stores ratio information indicating a ratio of an average figure of a person;
With
When the difference image and the foot pattern image representing the feature of the person's foot cannot be matched, the contact position specifying unit is indicated by the ratio information in the size of the other place of the person who has succeeded in matching. The monitoring system is characterized in that the ground contact position is specified by applying the ratio. The monitoring system according to claim 3, further comprising:
A ratio storage unit that stores ratio information indicating a ratio between the average face size of the person and the size of the whole body;
With
The grounding position specifying unit is
A face extraction unit that performs face recognition processing on the monitoring image and acquires a face size;
When the difference image and the foot pattern image representing the feature of the person's foot cannot be matched, the ratio indicated by the ratio information is applied to the size of the face extracted by the face extraction unit, A monitoring system characterized by specifying a grounding position.

Images