JP2005027085A - Image pick-up system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To lead out the number of persons passing in a digitizing region when one image pick-up device cannot photograph the digitizing region. <P>SOLUTION: An image pick-up system 10 has a plurality of cameras 3. The cameras 3 have digitizing modes in which only the digitizing regions allocated to the camera, are photographed by the camera, and monitor modes in which regions other than the digitizing regions are also photographed, as operating modes respectively. A digitizer 42 for a server device 2 leads out the number of the persons passing in the digitizing regions on the basis of images gained in the cameras 3. When one camera 3 transfers to the monitor mode, the camera 3 cannot photograph the digitizing region borne by the camera. Consequently, a camera control unit 41 makes the concerned digitizing region photograph to other cameras 3. Accordingly, the digitizing section 42 can lead out the number of the persons passing in the digitizing region. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for counting the number of passing objects.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there has been proposed an imaging system that counts the number of counted objects such as a person passing through a counting target area based on an image obtained by shooting a predetermined counting target area such as a passage.
[0003]
For example, an imaging system that captures a region to be monitored by an imaging device at a predetermined time period, extracts a human image from a captured image based on a background difference, and counts the number of passing people based on the extracted human image is known. (For example, refer to Patent Document 1).
[0004]
[Patent Document 1]
JP 2002-74371 A
[0005]
[Problems to be solved by the invention]
In recent years, it has been proposed to introduce an imaging system that counts the number of people passing through a store such as a supermarket or a convenience store, and obtain a customer's movement route and number of movements in the store as part of marketing research. On the other hand, usually in such stores, an imaging system including an imaging device (surveillance camera) that acquires images for observation by traveling around the store is arranged for the purpose of crime prevention. When two types of imaging systems for counting and observation are introduced in this way, problems such as a reduction in installation space and an increase in introduction cost arise.
[0006]
In order to solve this, it is conceivable that the imaging device can switch and acquire both images for counting and observation, and one imaging system can be used for both counting and observation. However, when the imaging device acquires an image for counting, it is necessary to fixedly shoot only a predetermined counting target area for counting, but when the imaging device acquires an image for observation. It is necessary to change the shooting target. Accordingly, when the imaging device acquires an image for observation, the imaging device cannot capture the counting target area in a fixed manner, and it is impossible to count the number of passages of persons with respect to the counting target area. Become. For this reason, one imaging system could not be used appropriately for both counting and observation purposes.
[0007]
The present invention has been made in view of the above problems, and an imaging system capable of deriving the number of passages of the counting object in the counting target area even when the imaging apparatus cannot capture the counting target area. The purpose is to provide.
[0008]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the invention of claim 1 is an imaging system based on a plurality of imaging devices that divide and shoot a plurality of count target areas and images obtained by the plurality of imaging devices. Derivation means for deriving the number of passages of the counting object in each of the plurality of counting target areas, and one imaging device of the plurality of imaging devices is in charge of photographing by the one imaging device When it is no longer possible to capture the first area, which is the counting target area, the derivation unit determines whether the counting target in the first area is based on an image acquired by an imaging apparatus different from the one imaging apparatus. Deriving the number of passes.
[0009]
The invention according to claim 2 is the imaging system according to claim 1, further comprising three or more imaging devices as the plurality of imaging devices, and control means for controlling operations of the plurality of imaging devices. When the one imaging device cannot photograph the first region, the control unit can photograph the first region from an imaging device other than the one imaging device among the plurality of imaging devices. And the selected imaging device is caused to photograph the first area.
[0010]
The invention according to claim 3 is the imaging system according to claim 1, wherein when the one imaging device cannot photograph the first area, the derivation means passes between the first area and the first area. Based on the image obtained by photographing the second area where there is no branch point, the number of passages of the counted object in the second area is derived, and based on the number of passages of the counted object in the second area. Thus, the number of passing objects to be counted in the first region is derived.
[0011]
According to a fourth aspect of the present invention, in the imaging system according to the first aspect, the first region is a single entrance / exit region related to a single passage branch point, and the single imaging device is the first region. When it becomes impossible to shoot the image, the derivation means, based on the images obtained by shooting the plurality of second areas as the entrance / exit areas excluding the first area related to the one passage branch point, Deriving the number of passages of the counting object in each of the plurality of second regions, and determining the number of passages of the counting object in the first region based on the number of passages of the counting object in each of the plurality of second regions. To derive.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following, an imaging system arranged in a store such as a supermarket or a convenience store will be described as an example.
[0013]
<1. First Embodiment>
<1-1. Configuration>
FIG. 1 is a diagram showing a schematic configuration of an imaging system 10 according to the first embodiment of the present invention. As shown in FIG. 1, the imaging system 10 includes a plurality of cameras 3 that are imaging devices that capture images and acquire images, and a server device 2 that controls the plurality of cameras 3. The plurality of cameras 3 are arranged on the ceiling of the sales area 100 for placing products and selling the products to customers, and the server device 2 is arranged in a store management room or the like. The number of cameras 3 is not particularly limited as long as it is plural.
[0014]
Each of the plurality of cameras 3 and the server device 2 is connected to a transmission cable 4 such as a LAN cable so that data communication can be performed between the devices. The data communication method is not limited to the wired method, and may be a wireless method.
[0015]
A sensor 5 is disposed at a predetermined position in the sales area 100. The sensor 5 detects shoplifting (carrying out an unsettled product out of a predetermined range) by detecting a tag attached to each unsettled product in the store. When the sensor 5 detects the tag, it generates a signal, and inputs the generated signal to the server device 2 via the transmission cable 51. In FIG. 1, only one sensor 5 is shown, but a plurality of sensors 5 may be provided.
[0016]
In the passage of the sales area 100, a plurality of count target areas serving as areas for counting the number of passing persons are set in advance. When the imaging system 10 is in a normal state, a plurality of cameras 3 shoot while sharing a plurality of counting target areas, and the number of persons passing through the counting target area is counted based on a counting image obtained by the shooting. The
When shoplifting is detected by the sensor 5, one camera 3 capable of shooting the area where shoplifting is detected is selected, and the selected camera 3 captures the area and obtains an image for observation. . For this reason, each of the plurality of cameras 3 includes a counting mode for acquiring a counting image and a monitoring mode for acquiring an observation image as its operation mode.
[0017]
The plurality of cameras 3 have the same configuration. FIG. 2 is a diagram illustrating a schematic configuration of the camera 3. As shown in FIG. 2, the camera 3 mainly includes a photographing lens 34 that forms an incident light, a CCD 35 that photoelectrically converts the formed light image into a signal charge, a signal processing circuit 36 that processes an output signal of the CCD 35, and an image. An image memory 37 temporarily stored, a communication unit 38 for communicating with the server device 2, a lens driving unit 39 for driving members of the photographing lens 34, and a control unit 30 for controlling the entire apparatus are provided.
[0018]
The photographing lens 34 can change the photographing magnification (zoom) and the optical axis position (pan, tilt). The lens driving unit 39 drives a member of the photographing lens 34 based on a signal from the control unit 30 and changes the photographing magnification and the optical axis position.
[0019]
The signal processing circuit 36 has an A / D converter. The analog signal charge output from the CCD 35 is converted into a digital image by the signal processing circuit 36 and then stored in the image memory 37. The image stored in the image memory 37 is transmitted to the server device 2 by the communication unit 38 based on the control of the control unit 30.
[0020]
The control unit 30 includes a microcomputer. Specifically, it includes a CPU 31 that performs various calculations, a RAM 32 that is a work area for the calculations, and a ROM 33 that records a control program. Various control functions of the control unit 30 are realized by the CPU 31 performing arithmetic processing in accordance with a control program recorded in the ROM 33.
[0021]
Such a control function of the control unit 30 includes a function of controlling each unit so as to operate in accordance with the operation mode. For example, in the monitoring mode, the control unit 30 performs control to change the shooting magnification and the optical axis position of the shooting lens 34 so that shooting with tracking a specific subject or traveling shooting in a store can be performed. On the other hand, in the counting mode, the control unit 30 performs control to fix the photographing magnification and the optical axis position of the photographing lens 34 so that a predetermined counting target region can be photographed. The control unit 30 also has a function of controlling each unit so as to perform an operation according to a signal from the server device 2.
[0022]
FIG. 3 is a diagram illustrating a basic configuration of the server device 2. The hardware configuration of the server device 2 is almost a general computer configuration. That is, the server device 2 includes a CPU 21 that performs various arithmetic processes, a ROM 22 that stores basic programs, a RAM 23 that serves as a work area for arithmetic processes, a hard disk 24 that stores programs and various data files, and various displays. A display 25 for performing and an input unit 26 including a keyboard and a mouse are provided. The server device 2 further includes a communication unit 27 for communicating with the camera 3 and a sensor input unit 28 for inputting a signal from the sensor 5.
[0023]
A dedicated program is stored in advance in the hard disk 24 of the server device 2, and various functions as the server device 2 are realized by the CPU 21 performing arithmetic processing according to the program.
[0024]
FIG. 4 is a diagram illustrating a functional configuration of the imaging system 10 including functions of the server device 2 realized by the CPU 21 performing arithmetic processing according to a program.
4, the camera control unit 41 and the counting unit 42 are realized by the CPU 21 performing arithmetic processing according to a program.
[0025]
The camera control unit 41 controls each operation of the plurality of cameras 3. The camera control unit 41 can perform various controls such as changing the operation mode and changing the optical axis position of the photographing lens 34 by transmitting a predetermined signal to the camera 3 via the communication unit 27. . The counting unit 42 counts the number of passing persons in each of the plurality of counting target areas based on the images obtained by the plurality of cameras 3.
[0026]
Next, a specific example of a plurality of count target areas set in the sales area 100 will be described. FIG. 5 is a diagram showing a state from the upper part of the area 101 which is a part of the sales area 100. Hereinafter, the XY plane coordinate system shown in FIG. 5 is used when indicating the direction and orientation.
[0027]
As shown in FIG. 5, in the area 101, a plurality of product shelves Sh for placing products are arranged, and a space between the plurality of product shelves Sh is a passage P through which a person moves. . And the count object area | region 60 is set on the ground surface of such a channel | path P. As shown in FIG. In the area 101, five count target areas 60 (A to E) are set.
[0028]
The count target area A and the count target area B are set on the same path P1 where there is no path branch point (hereinafter referred to as “path branch point”). More specifically, the counting target area A is set to the X axis minus end of the passage P1, and the counting target area B is set to the X axis plus end of the passage P1.
[0029]
Further, the X axis + side of the counting target region B is a passage branch point Cp of the passage P, and the counting target region B is one entry / exit region related to the passage branching point Cp (on the passage leading to the passage branching point). It is also a region close to the passage branch point). The count target areas D to E are also entrance / exit areas related to the passage branch point Cp. Specifically, the count target area C is on the path P2 on the X axis + side of the path branch point Cp, the count target area D is on the path P3 on the Y axis + side of the path branch point Cp, and the count target area E is a path branch. They are set on the passage P4 on the Y axis minus side of the point Cp.
[0030]
In the area 101, five cameras 3 are arranged. Ranges 50a to 50e in the figure respectively indicate ranges to be imaged when these five cameras 3a to 3e are in the counting mode. That is, the camera 3a is responsible for photographing the counting target area A, the camera 3b is responsible for photographing the counting target area B, the camera 3c is responsible for photographing the counting target area C, and the camera 3d is photographing the counting target area D. The camera 3e is in charge of photographing the counting target area E. As described above, in the present embodiment, one camera 3 is in charge of photographing one counting target region 60.
[0031]
<1-2. Processing details>
Next, processing contents of the imaging system 10 will be described with reference to FIG.
In the imaging system 10, when the operation starts, the operation mode of all the cameras 3 is set to the count mode. Then, the counting target region 60 in charge of each camera 3 is photographed at a predetermined time period (for example, 1/30 second period), and an image acquired at the predetermined time period is transmitted to the server device 2. The
[0032]
Thereby, in the server apparatus 2, the some image which shows each state of the some count object area | region 60 in a shop is received by the communication part 27 by a predetermined | prescribed time period. The received image is stored in the RAM 22. The counting unit 42 obtains the number of passing persons in each of the plurality of counting target areas 60 based on the image stored in the RAM 22 in this way.
[0033]
FIG. 6 shows an example of an image 70 obtained by one camera 3. As shown in the figure, the image 70 includes an image 74 showing the product shelf Sh, an image 75 showing the passage P, an image showing a person (hereinafter referred to as “person image”) 79, and the like. An area (hereinafter referred to as “in-image target area”) 73 corresponding to the count target area 60 is also included in the image 70.
[0034]
The counting unit 42 extracts only a person image from such an image 70 based on the background difference, and counts the number of passing persons based on the extracted person image. That is, first, a difference between the image 70 and a background image 71 (see FIG. 7) showing only a background (product shelf image 75, passage image 75) other than a person is taken, and a difference image 72 (see FIG. 8) is taken. .) Is generated.
The background image 71 is stored in the hard disk 24 in advance. As shown in FIG. 8, only the person image 79 is extracted from the difference image 72.
[0035]
Subsequently, based on the positional relationship between the person image 79 and the in-image target area 73 in the difference image 72, it is determined whether or not a person exists in the counting target area 60. When a person is present in the counting target area 60, the number of passages related to the counting target area 60 is increased by one. Data indicating which area in the difference image 72 (image 70) is the in-image target area 73 is stored in the hard disk 24 in advance. In this way, the number of passing persons in the counting target area 60 is counted.
[0036]
Further, in the process of counting the persons, the traveling direction of the person is also determined. Specifically, the person image 79 in the difference image 72 obtained most recently is compared with the person image 79 in the difference image 72 obtained in the past, and the moving direction of the person image 79 in the image is determined. Then, based on the moving direction in the image of the person image 79, the traveling direction of the person on the actual passage P is determined. The correspondence between the orientation in the difference image 72 and the orientation on the actual passage P is stored in the hard disk 24 in advance. In this way, the counting unit 42 determines the traveling direction of the person and counts the number of passing persons for each traveling direction.
[0037]
Such a calculation is performed for all images obtained from the camera 3, whereby the number of passing persons in each counting target area 60 in the store is obtained for each traveling direction. Then, the number of passages of the person obtained for each traveling direction in this way is totaled every predetermined time interval (for example, 5 seconds), and is collected in a result table T having a table format as shown in FIG. Hereinafter, the total time is referred to as “total time”.
[0038]
As shown in FIG. 9, in the result table T, each field (column) F is one counting direction region 60 (A, B, C, D,...) In one traveling direction (X +, X−, Y +, Y-) respectively. Each record (row) corresponds to the total time (Te1, Te2,...). That is, one record is generated every time one total time is completed. In this way, the number of passing persons (Nap1,..., Nem7) is counted for each traveling direction of the counting target area 60 and for each counting time.
[0039]
The above is the processing contents at the normal time of the imaging system 10, but when shoplifting is detected by the sensor 5 when such processing is performed (when a tag attached to a product is detected), The operation mode of the camera 3 is shifted from the counting mode to the monitoring mode. That is, the operation mode of one camera 3 is monitored from the counting mode when an event that satisfies a predetermined reference condition for shifting the operation mode (in this embodiment, detection of a tag by the sensor 5) occurs. It will be transferred to the mode.
[0040]
More specifically, when a tag is detected by the sensor 5, a predetermined signal is input from the sensor 5 to the server device 2. In response to this, the sensor input unit 28 transmits the fact and the position of the sensor 5 to the camera control unit 41. Subsequently, the camera control unit 41 selects one camera 3 that is closest to the position of the sensor 5 and capable of photographing the position of the sensor 5 based on the position of the sensor 5. Next, for the selected camera 3, the camera control unit 41 causes the operation mode to shift from the counting mode to the monitoring mode, and a signal indicating the position to be photographed (the position of the sensor 5 where shoplifting has occurred). Is sent.
[0041]
The camera 3 to which these signals are transmitted shifts the operation mode from the count mode to the monitoring mode. Then, after changing the optical axis position and the photographing magnification of the photographing lens 34 so that the position to be photographed can be photographed appropriately, processing for photographing the position at a predetermined time period is started. In addition, when a person is present in the captured image, the camera 3 derives the position of the person by a known technique such as time difference or background difference, and moves the optical axis position so as to track the person. Let The image acquired in this way is transmitted to the server device 2 and displayed on the display 25 as an image for observation and recorded on the hard disk 24. As a result, an image showing the person who has shoplifted is acquired.
[0042]
When one camera 3 shifts to the monitoring mode in this way, the camera (hereinafter referred to as “transition camera”) is referred to as a count target area (hereinafter referred to as “first attention area”) for which the device itself is responsible for shooting. ) 60 cannot be captured, and as a result, the counting unit 42 cannot count the number of passing persons in the first region of interest.
[0043]
Therefore, in the present embodiment, when one camera 3 shifts to the monitoring mode, the camera control unit 41 selects one camera 3 different from the transition camera that can capture the first attention area. Which camera 3 can shoot each counting target area 60 is stored in advance in the hard disk 24 as data. One camera (hereinafter referred to as “support camera”) 3 capable of photographing the first region of interest is selected by referring to such data.
[0044]
That is, when the total number of cameras 3 is two, only one camera remains as a camera other than the transition camera that has entered the monitoring mode, so that the camera is used as a substitute for the counting function of the transition camera. However, if the total number of cameras 3 is three or more, there are two or more cameras other than the transition camera. is there. On the other hand, depending on which transition camera is used, which other camera is suitable as a support camera differs, so for each camera, which other camera is the support camera when that camera becomes the transition camera. It is specified in advance as to whether it is to be used as data (for example, data in a table format).
[0045]
When one camera 3 capable of photographing the first region of interest is selected, the camera control unit 41 transmits a signal instructing photographing of the first region of interest to the camera 3. Thereby, the first attention area is photographed by the support camera. The acquired image is transmitted to the server device 2. The counting unit 42 derives the number of passages of the person in the first region of interest based on the image thus captured by the support camera.
[0046]
Hereinafter, this process will be specifically described by taking the area 101 shown in FIG. 5 as an example. Here, it is assumed that the operation mode of the camera 3b in charge of photographing the counting target area B is shifted to the monitoring mode, and the camera 3b can no longer photograph the counting target area B. The counting area B can be taken by the camera 3a. Therefore, the camera control unit 41 selects the camera 3a as a support camera. Then, a signal for instructing photographing of the counting target region B is transmitted to the camera 3a, and the optical axis position and photographing magnification of the photographing lens 34 of the camera 3a are changed. As a result, as shown in FIG. 10, the counting target area A together with the counting target area A for which the camera 3a is in charge of shooting is included in the range 50a captured by the camera 3a.
[0047]
FIG. 11 is a diagram showing an example of an image acquired by the camera 3a in this way. As shown in the figure, the image 70 includes an image 74 indicating the product shelf Sh, an image 75 indicating the passage P, and a person image 79, and an in-image target area 73a corresponding to the counting target area A, In addition, an in-image target area 73b corresponding to the counting target area B is included. The counting unit 42 derives the number of passing persons in both the counting target area A and the counting target area B based on such an image. In this way, even when the camera 3b is unable to shoot the counting target area B, the number of persons passing through the counting target area B can be counted based on the image acquired by the camera 3a. It becomes.
[0048]
In general, other cameras that are installed so as to be able to image not only their own counting target area but also the counting target area of the transition camera are selected as support cameras, but when there are a plurality of such cameras, It is preferable in terms of imaging accuracy (and thus counting accuracy) that the camera closest to the counting target area of the transfer camera is designated as the support camera. In addition, when the counting target area of the transfer camera cannot be covered by the angle of view of one support camera, a plurality of support cameras can be cooperated as an alternative to the transfer camera.
[0049]
As described above, in the imaging system 10 of the present embodiment, even when one camera 3 shifts to the monitoring mode and the count target area in charge of shooting cannot be shot, the count target area Another camera 3 capable of capturing the image is selected as a support camera, and the count target area is captured by the support camera. In other words, it is possible to back up the surveillance mode when switching to multiple cameras by providing a possibility of replacement (redundancy) in at least a part of the imageable range without completely isolating the imageable range. Will be. Thereby, it is possible to derive the number of passing persons in the counting target area handled by the camera 3 that has shifted to the monitoring mode. Therefore, the imaging system 10 of the present embodiment can be appropriately used for both counting and observation applications.
[0050]
<2. Second Embodiment>
Next, a second embodiment of the present invention will be described. The configuration of the imaging system 10 of the present embodiment is the same as that of the first embodiment. In the first embodiment, when one camera 3 cannot photograph the first region of interest, the first attention is based on an image obtained by photographing the first region of interest with the other camera 3. In the present embodiment, the number of passages in the first region of interest is derived from the number of passages of persons in the counting target region 60 different from the first region of interest. Yes.
[0051]
Here, as in the first embodiment, it is assumed that the camera 3b shifts to the monitoring mode in the area 101 shown in FIG. FIG. 12 is a diagram showing a state of the area 101 in this case. As shown in FIG. 12, the number of passages in the X axis + direction of the counting target region B in a certain counting time is Nbp, and the number of passages in the X axis-direction is Nbm. Further, the number of passages in the X axis + direction of the counting target region B in a certain counting time is Nap, and the number of passages in the X axis-direction is Nam.
[0052]
As shown in the figure, the count target area A and the count target area B are areas on the same path P1, and there is no path branch point between these areas. In general, it can be considered that the time during which a person moves between the counting target area A and the counting target area B (hereinafter, the symbol Δt is used) is substantially constant. Therefore, a person who has passed through the counting target area A in the X axis + direction must pass through the counting target area B after time Δt, and conversely, a person who has passed through the counting target area B in the X axis−direction It is considered that it always passes through the counting target area A after Δt.
[0053]
Therefore, the passage number Nbp in the X axis + direction of the counting target region B at a certain counting time Tex matches the passing number Nap in the X axis + direction of the counting target region A in the counting time that is Δt before the counting time Tex. Then it can be considered. On the other hand, when the counting target area B at a certain counting time Tex is in the X-axis-direction passing number Nbm coincides with the counting number area Am of the counting target area A at the counting time At ahead of the counting time Tex. Can be considered. That is, the number of passages in the count target region B can be derived based on the number of passes of the count target region A in which no passage branch point exists between the count target region B.
[0054]
The counting unit 42 of the present embodiment uses such a principle, and when one camera 3 shifts to the monitoring mode and imaging of the first attention area becomes impossible, the counting section 42 is between the first attention area. The number of passages in the first region of interest is derived from the number of passages of persons in the counting target region 60 on the same passage where there is no passage branch point.
[0055]
Specifically, when the operation mode of one camera 3 shifts to the monitoring mode, first, an image for counting is acquired from a camera 3 other than the camera 3, and the first region of interest is based on the counting image. The number of passing persons in each counting target region 60 other than is derived.
[0056]
Then, one counting target region on the same passage where no passage branch point exists between the first attention region and the first attention region is selected as the “second attention region”. The path P on which each count target area 60 exists is stored in advance in the hard disk 24 as data. The second region of interest is selected with reference to such data. The number of passing persons in the second region of interest has been derived in the above calculation.
[0057]
Subsequently, the movement time Δt of the person between the first attention area and the second attention area is acquired.
This movement time Δt is also stored in the hard disk in advance. The number of passages of the first attention area in the direction from the second attention area toward the first attention area at a certain aggregation time Tes is the second attention area at the aggregation time corresponding to the movement time Δt before the aggregation time Tes. The same number of passes in the same direction is used. On the other hand, the number of passages of the first attention area in the direction from the first attention area to the second attention area at a certain aggregation time Tes is the second attention area at the aggregation time corresponding to the movement time Δt after the aggregation time Tes. The number of passages in the same direction will be used.
[0058]
As described above, in the present embodiment, the number of passages of the second region of interest in which there is no passage branch point between the first region of interest and the first region of interest is derived based on the number of passages of the second region of interest. The number of passes in the region is derived. Therefore, even in the present embodiment, even when one camera 3 shifts to the monitoring mode and the camera 3 cannot capture the counting target area (first attention area) for which imaging is performed, The number of passages of a person in one attention area can be derived.
[0059]
Further, in the present embodiment, it is not necessary for the camera 3 different from the camera 3 that has shifted to the monitoring mode to change the optical axis direction or the imaging magnification (imaging position). Therefore, the technique of the present embodiment can be applied to an imaging system including a camera that does not have a function of changing the optical axis direction or the imaging magnification.
[0060]
If there is no counting target area where there is no passage branch point between the first attention area, an area where no passage branch point exists between the first attention area is set as the calculation area, The area is photographed by another camera 3, the number of passes in the calculation area is derived from the image obtained by the camera 3, and the number of passes of the first attention area is derived based on the number of passes in the calculation area. You may do it.
[0061]
For example, as shown in FIG. 13, it is assumed that the count target area A does not exist in the passage P1 of the area 101 that is substantially the same as that in FIG. In this area 101, it is assumed that the counting target area F exists on the X-axis-side of the passage branch point Cpp, and the camera 3f that captures the range 50f including the counting target area F in the normal state.
[0062]
In such an area 101, it is assumed that the camera 3b shifts to the monitoring mode and the camera 3b can no longer capture the count target area B. In this case, there is no counting target area where there is no passage branch point with the counting target area B. For this reason, as shown in FIG. 14, a calculation area G that is an area where no passage branch point exists with respect to the count target area B is set. Then, the direction of the optical axis and the shooting magnification of the camera 3f are changed, and the calculation area G is shot by the camera 3f, and the number of passes in the calculation area G is derived based on the image acquired by the camera 3f. Thereby, the number of passages in the counting target region B can be derived by the same method as described above based on the number of passages in the calculation region G.
[0063]
<3. Third Embodiment>
Next, a third embodiment of the present invention will be described. The configuration of the imaging system 10 of the present embodiment is the same as that of the first embodiment. Also in the present embodiment, as in the second embodiment, the number of passages in the first region of interest is derived from the number of passages of people in the counting target region 60 different from the first region of interest.
[0064]
Here, as in the above embodiment, in the area 101, as shown in FIG. 12, it is assumed that the camera 3b shifts to the monitoring mode and the camera 3b can no longer capture the count target area B. As shown in FIG. 12, the number of passages in the X axis + direction of the counting target region B in a certain counting time is Nbp, and the number of passages in the X axis-direction is Nbm. In addition, the number of passages in the X axis + direction of the counting target region C at a certain counting time is Ncp, the number of passages in the X axis − direction is Ncm, and the number of passages in the Y axis + direction of the counting target region D is Ndp, Y axis. The number of passages in the − direction is Ndm, the number of passages in the Y axis + direction of the count target region E is Nep, and the number of passages in the Y axis − direction is Nem.
[0065]
A person who passes through the counting target region B in the X axis + direction and enters the passage branch point Cp is considered to normally pass through any of the counting target regions C, D, E and exit from the passage branch point Cp. In general, it can be considered that the time from when the person enters the passage branch point Cp until he leaves (hereinafter, the symbol ΔT is used) is substantially constant. Accordingly, the number Nbp of people passing through the counting target area B and entering the passage branch point at a certain counting time Tex passes through each of the counting target areas C, D, E at the counting time after ΔT from the counting time Tex. It will be included in the added value of the numbers Ncp, Ndp, and Nem from the passage branch point Cp. Further, the added value of Ncp, Ndp, and Nem includes the numbers Ncm, Ndm, and Nep that pass through each of the count target areas C, D, and E and enter from the passage branch point Cp at the total time Tex. For this reason, a value obtained by subtracting Ncm, Ndm, and Nep at the total time Tex from the addition value of Ncp, Ndp, and Nem at the total time after ΔT from the total time Tex passes through the count target region B and branches off the passage. It can be considered that the number of people entering the point is Nbp. On the contrary, a value obtained by subtracting Ncp, Ndp, and Nem at the total time Tex from the addition value of Ncm, Ndm, and Nep at the total time ΔT before the total time Tex is counted at a certain total time Tex. It can be considered that the number of people Nbm passes through the target area B and exits from the passage branch point.
[0066]
The counting unit 42 of the present embodiment uses such a principle, and when one camera 3 shifts to the monitoring mode and imaging of the first attention area becomes impossible, the path branch except for the first attention area Based on the number of passes in the entry / exit area related to the point, the number of passes of the first region of interest is derived.
[0067]
Specifically, when the operation mode of one camera 3 shifts to the monitoring mode, first, an image for counting is acquired from a camera 3 other than the camera 3, and the first region of interest is based on the counting image. The number of passing persons in each counting target region 60 other than is derived.
[0068]
Then, a path branch point close to the first attention area is noticed, and a plurality of entry / exit areas related to the path branch point excluding the first attention area are selected as “second attention areas”. Which path branching point each counting target area 60 is close to is stored in advance in the hard disk 24 as data. The second region of interest is selected with reference to such data. The number of passing persons in the plurality of second attention areas has been derived in the above calculation.
[0069]
Subsequently, the movement time ΔT of the person passing through the path branch point where the first region of interest is close is acquired. This movement time ΔT is also stored in the hard disk in advance. Then, the number of passages of the first region of interest in the direction of entering the passage branch point at a certain total time Tes comes out from the passage branch points of the plurality of second target regions at the total time corresponding to the movement time ΔT after the total time Tes. A value obtained by subtracting the addition value of the number of passages in the direction entering the passage branch point of the plurality of second attention areas at the total time Tes from the addition value of the number of passages in the direction is used. On the other hand, the number of passages of the first region of interest in the direction exiting from the passage branch point at a certain aggregation time Tes is to the passage branch points of a plurality of second attention regions at the aggregation time corresponding to the movement time ΔT before the aggregation time Tes. A value obtained by subtracting the addition value of the number of passages in the direction exiting from the passage branch point of the plurality of second attention areas at the total time Tes from the addition value of the number of passages in the entering direction is used.
[0070]
As described above, in the present embodiment, the number of passages of a plurality of second attention areas which are entry / exit areas related to a passage branch point where the first attention area is close is derived. Then, the number of passes in the first region of interest is derived based on the number of passes of the second region of interest. Therefore, even in the present embodiment, even when one camera 3 shifts to the monitoring mode and the camera 3 cannot capture the counting target area (first attention area) for which imaging is performed, The number of passages of a person in one attention area can be derived.
[0071]
Also in the present embodiment, it is not necessary for the camera 3 different from the camera 3 that has shifted to the monitoring mode to change the optical axis direction or imaging magnification (imaging position). Therefore, the technique of the present embodiment can be applied to an imaging system including a camera that does not have a function of changing the optical axis direction or the imaging magnification.
[0072]
In the present embodiment, additional information such as the color, size, speed, etc. of the person's clothes may be obtained in the calculation, and these information may be referred to when the number of passages is derived.
[0073]
As described above, in the second and third embodiments, generally, there are a plurality of entrances / exits in a specific space range (passage space in the second embodiment, passage branch point in the third embodiment), In the case where all of these multiple entrances can be imaged by any camera, when one camera provided corresponding to one entrance becomes a transition camera, the other cameras By using this function, it is possible to estimate the number of passages that the transition camera would have counted. This is generally the case where the number of counted objects is preserved (that is, when the counted objects are neither generated nor disappeared in the middle) and the time required for the counted objects to pass through the spatial range. The system configuration is particularly suitable for cases where an upper limit can be assumed.
[0074]
<4. Fourth Embodiment>
In the above-described embodiment, one camera 3 is in charge of photographing one counting target region 60 in normal times. Of course, one camera 3 is responsible for photographing a plurality of counting target regions 60. It may be like this.
[0075]
However, in this case, when the operation mode of one camera 3 shifts to the monitoring mode, it is necessary to derive the number of passages in each of the plurality of count target areas 60 that the camera 3 is responsible for shooting. At this time, the number of passages of each of the plurality of count target regions 60 may be derived by combining the methods of the first to third embodiments. Specific examples will be described below.
[0076]
FIG. 15 is a diagram illustrating an example of a plurality of count target areas set in the sales area 100 of a certain store. As shown in the figure, in the sales area 100, a plurality of product shelves Sh are placed, and 13 count target areas (H to T) are set.
[0077]
In the sales area 100, four cameras 3h, 3j, 3n, 3r are arranged. Ranges 50h, 50j, 50n, and 50r in the figure indicate ranges that are imaged when the four cameras 3h, 3j, 3n, and 3r are in the counting mode, respectively. That is, the camera 3h is in charge of shooting the counting target area H near the store entrance / exit En, the camera 3j is in charge of shooting the four counting target areas IL, and the camera 3n is in the four counting target areas MP. The camera 3r is in charge of shooting, and the camera 3r is in charge of shooting of the four counting target areas Q to T.
[0078]
FIG. 16 shows another camera 3h, 3n, 3r when the camera 3j shifts to the monitoring mode in the sales area 100 and the camera 3j cannot photograph the count target areas I, J, K, L. The example of each imaging | photography range 50h, 50n, 50r is shown.
[0079]
As shown in the figure, the count target areas I and J are directly photographed by the camera 3h. Therefore, the number of passages in the count target areas I and J can be derived based on the image acquired by the camera 3h by the same method as in the first embodiment.
[0080]
In addition, a calculation area U that is an area where no passage branch point exists between the calculation target area L and the calculation area U is photographed by the camera 3n. Therefore, the number of passages of the count target region L can be derived based on the number of passages of the calculation region U by a method similar to that of the second embodiment.
[0081]
The count target area K is one entry / exit area of the passage branch point Cpx. Then, the number of passages of the plurality of count target areas J, N, O, R, and S that are entrance / exit areas other than the count target area K of the passage branch point Cpx is acquired by any of the cameras 3h, 3n, and 3r, respectively. It can be obtained directly based on the obtained image. Therefore, the number of passages of the count target region K can be derived based on the number of passes of the count target regions J, N, O, R, and S by the same method as in the third embodiment.
[0082]
As described above, the counting target areas I, J, K, and L that the camera 3j takes charge of can be derived by combining the techniques of the first to third embodiments. It becomes.
[0083]
<5. Modification>
Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications are possible.
[0084]
For example, in the above-described embodiment, the sensor 5 that generates a signal that responds when the operation mode of the camera 3 shifts from the counting mode to the monitoring mode detects shoplifting, but a disaster such as a fire or an earthquake. Or a device that detects an air condition (more generally, a room condition) such as temperature or humidity.
[0085]
In addition, the reference condition for shifting the operation mode of the camera 3 is not limited to the generation of the signal of the sensor 5, and may be made when other conditions are satisfied. For example, when an abnormal object or person is found in the image acquired by the camera 3, the operation mode of the camera 3 may be shifted.
[0086]
In the above embodiment, the object to be counted is a person. However, any object that moves in a predetermined area to be counted, such as an object such as luggage, a vehicle such as an automobile, or an animal, may be used. May be.
[0087]
【The invention's effect】
As described above, according to the first to fourth aspects of the present invention, even when one imaging apparatus is unable to capture the counting target area for which the apparatus is in charge of imaging, Based on the acquired image, it is possible to derive the number of passing objects to be counted in the counting target area in charge of one imaging apparatus.
[0088]
In particular, according to the invention of claim 2, since the imaging device capable of photographing the first area directly photographs the first area, the number of passing objects to be counted in the first area can be accurately derived. it can.
[0089]
In particular, according to the third and fourth aspects of the present invention, it is possible to derive the number of passing objects to be counted in the first region without changing the imaging position of the imaging device.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a schematic configuration of an imaging system.
FIG. 2 is a diagram illustrating a schematic configuration of a camera.
FIG. 3 is a diagram illustrating a basic configuration of a server device.
FIG. 4 is a diagram illustrating a functional configuration of the imaging system.
FIG. 5 is a diagram illustrating a state of a partial area of a sales area.
FIG. 6 is a diagram illustrating an example of an image obtained by a camera.
FIG. 7 is a diagram illustrating an example of a background image.
FIG. 8 is a diagram illustrating an example of a difference image.
FIG. 9 shows a result table.
FIG. 10 is a diagram illustrating a state of a partial area of a sales area.
FIG. 11 is a diagram illustrating an example of an image obtained by a camera.
FIG. 12 is a diagram illustrating a state of a partial area of a sales area.
FIG. 13 is a diagram illustrating a state of a partial area of a sales area.
FIG. 14 is a diagram showing a state of a partial area of a sales area.
FIG. 15 is a diagram showing a state of a sales area.
FIG. 16 is a diagram showing a state of a sales area.
[Explanation of symbols]
2 Server device
3 Camera
5 Sensor
10 Imaging system
41 Camera control unit
42 counter
60 Target area

Claims (4)

An imaging system,
A plurality of imaging devices that divide and shoot a plurality of counting target areas;
Derivation means for deriving the number of passing objects to be counted in each of the plurality of counting target regions based on images obtained by the plurality of imaging devices,
When one imaging device out of the plurality of imaging devices cannot photograph the first area that is the counting target area for which the one imaging apparatus is in charge of photographing,
The derivation means derives the number of passages of the counted object in the first region based on an image acquired by an imaging device different from the one imaging device. The imaging system according to claim 1,
While comprising three or more imaging devices as the plurality of imaging devices,
Control means for controlling operations of the plurality of imaging devices,
When the one imaging device cannot capture the first area,
The control means selects an imaging device capable of photographing the first region from imaging devices other than the one imaging device among the plurality of imaging devices, and causes the selected imaging device to photograph the first region. A characteristic imaging system. The imaging system according to claim 1,
When the one imaging device cannot photograph the first region, the derivation means includes:
Based on the image obtained by photographing the second area where there is no passage branch point between the first area and the first area, the number of passages of the counted object in the second area is derived,
An imaging system, wherein the number of passing objects to be counted in the first area is derived based on the number of passing objects to be counted in the second area. The imaging system according to claim 1,
The first region is one entrance / exit region related to one passage branch point,
When the one imaging device cannot photograph the first region, the derivation means includes:
Based on images obtained by photographing a plurality of second areas as entrance / exit areas excluding the first area related to the one passage branch point, the counting object in each of the plurality of second areas Deriving the number of passes,
An imaging system, wherein the number of passages of the counting object in the first region is derived based on the number of passages of the counting object in each of the plurality of second regions.

Images