JP2015139204A - Method for evaluating rat detection system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for evaluating a rat detection system capable of evaluating the detection function of the rat detection system installed in a facility.SOLUTION: In a state of moving a rat simulation body 200 in a rat detection target area E, the appearance determination of the rat simulation body 200 is stored by the processing of an imaging unit 20, an image processing unit 50 and a determination unit 70. In addition, apart from the appearance determination by a rat detection system 10, the existence locus of the rat simulation body 200 in the rat detection area is stored into an existence locus storage unit 130. The evaluation of the rat detection system 10 is granted by the comparison of the appearance determination of the rat simulation body 200 stored in a storage unit 60 with the existence locus of the rat simulation body 200.

Description

  The present invention relates to a method for evaluating a mouse detection system.

  Conventionally, in order to grasp the invasion route and movement route of a mouse into a facility such as a factory, the mouse detection is provided with a sensor or an imaging means for detecting a mouse such as a photoelectric type, an infrared type, or a video camera. Devices have been proposed (Patent Document 1, Patent Document 2, Patent Document 3).

  The mouse detection system disclosed in Patent Document 1 grasps not only the existence of a mouse but also a movement path and a behavior pattern of the mouse based on moving image data obtained by moving the mouse detection target area with a video camera. .

The mouse detection apparatus disclosed in Patent Document 2 is configured not to count a single mouse redundantly and to distinguish the detection of a mouse from the detection of a person.
The mouse detection apparatus of Patent Document 3 is configured to divide a detection area into a first area to a fourth area including an entrance / exit of a mouse by two boundary setting lines that cross each other. Moreover, the 1st, 2nd sensor which has a light projection part and a light-receiving part is arrange | positioned at the both sides of the two said boundary setting straight lines. Then, the mouse detection device sets the light from the first light projecting unit to be in the first region, and from the first region to the second and third regions based on the signals of the two sensors. Only when the movement of the mouse is determined, the detection by the sensor is enabled.

JP 2011-259376 A JP 08-182456 A JP 2004-129594 A

  By the way, conventionally, when such a mouse detection system or a mouse detection device is actually set in a facility, the system itself in the set facility or the detection function of the device has not been evaluated.

  For this reason, when these systems were installed and operated in the facility, mice were not detected even though mice appeared, or mice were not detected because no mice appeared. It is the present situation that we have not been able to confirm whether or not.

  An object of the present invention is to provide an evaluation method for a mouse detection system capable of evaluating a detection function of a mouse detection system installed in a facility.

  An evaluation method for a mouse detection system of the present invention includes an imaging unit that images a mouse detection target area, an image processing unit that extracts a focused image from an image captured by the imaging unit, and performs feature extraction of the focused image, In the evaluation method of a mouse detection system having a determination unit for determining the appearance of a mouse based on feature extraction, the imaging unit, the image in a state where a mouse simulated body is arranged or moved in the mouse detection target area The appearance determination of the mouse simulated body is stored by the processing of the processing unit and the determination unit, and the presence locus of the mouse simulated body in the mouse detection target area is stored separately from the appearance determination by the mouse detection system. The first step stored in the part, the appearance determination of the mouse simulated body stored in the existence locus storage unit, and the presence of the mouse simulated body By comparing the trace, in which a second step of applying the evaluation of the murine detection system.

  Further, the image processing unit uses a reference image obtained by imaging the mouse detection target area when the mouse simulated body does not exist as a background image, and the image taken by the imaging unit and the reference image by the background subtraction method. It is preferable to extract the image of interest.

In addition, it is preferable that the mouse simulation body includes at least one of a mouse eyeball reflector, a mouse simulation shape, and an infrared light emitter.
Moreover, when the said mouse simulation body is provided with the said infrared light-emitting body, it is preferable that the said imaging part is an infrared camera.

  When the mouse simulation body includes the reflector, the mouse detection system preferably includes an infrared light projecting unit that projects infrared light toward the mouse detection target area, and the imaging unit is an infrared camera. .

  The mouse simulation body includes an automatic traveling unit that performs automatic traveling, and a sensor that detects a parameter for calculating an existence locus of the mouse simulated body when traveling by the automatic traveling unit, the first step. Then, the presence locus of the mouse simulated body may be calculated based on the parameters.

  According to the evaluation method of the mouse detection system of the present invention, the detection function of the mouse detection system installed in the facility can be evaluated.

The conceptual diagram of the evaluation apparatus of the mouse | mouth detection system of one Embodiment. The conceptual diagram of a mouse simulation body. 1 is a block diagram of a mouse detection system. The flowchart which an evaluation apparatus performs. A background image captured by the imaging unit. An image of a mouse simulated body imaged by the imaging unit. The difference image of FIG. 6 and FIG.

Hereinafter, an evaluation device and an evaluation method for a mouse detection system according to an embodiment of the present invention will be described with reference to FIGS.
First, the mouse detection system 10 to be evaluated will be described.

  The mouse detection system 10 according to the present embodiment is arranged at a detection target location including, for example, a building such as a building, a factory, a house, an underground structure (tunnel) such as a tunnel, and the like. This is for obtaining or detecting the intrusion of the mouse itself.

  As shown in FIG. 1, the mouse detection system 10 includes an imaging unit 20 and a determination device 30 that performs image processing of an image signal captured by the imaging unit 20 and presence / absence determination processing of a mouse. The imaging unit 20 includes, for example, a CCD camera, a CMOS camera, an infrared camera, or the like. As shown in FIG. 1, the imaging range by the imaging unit 20 includes at least a mouse detection target area E to be detected. The mouse detection target area E includes, for example, the floor of the building, the floor of an underground structure, the ground, or a road. The light projector 40 gives a light amount necessary for imaging the mouse detection target area E. When photographing with visible light, it is preferable that the projector 40 is composed of a visible light illumination lamp or the like and illuminates the mouse detection target area E. Moreover, when image | photographing with an infrared camera, the light projector 40 is comprised from the infrared light projector, and illuminates the mouse | mouth detection target area E with infrared rays.

  As illustrated in FIG. 3, the determination device 30 extracts an image of interest from an image captured by the imaging unit 20, extracts an image from the image of interest, and extracts an appearance of a mouse based on the feature extraction of the image of interest. And a storage unit 60 for storing the determination result of the determination unit.

  The image processing unit 50 will be described. The image processing unit 50 includes a background generation unit 51, a background storage unit 52, a difference processing unit 53, a binarization processing unit 54, and a feature extraction processing unit 55. In the present embodiment, the image processing unit 50 performs feature extraction of the image of interest using the background difference method. Specifically, an image obtained by imaging the mouse detection target area E by the imaging unit 20 in a state where no mouse appears, is stored in the background storage unit 52 via the background generation unit 51 as a background image. The background generation method is generated by a background integration method in which images are sequentially replaced at a certain ratio by the background generation unit 51, or by a method of sequential averaging. The difference processing unit 53 performs a difference process between the background image generated by these methods and the one-frame image acquired by the imaging unit 20 for detecting a mouse, and acquires the difference image. The binarization processing unit 54 binarizes using the set threshold value. The pixel of the target image obtained by the binarization process is labeled by the feature extraction processing unit 55, and the feature extraction process is performed for each target image. The feature here is information obtained from the image of interest such as the area of the image of interest, the position of the center of gravity of the image of interest, and the end point.

  Based on the features extracted for each image of interest by the feature extraction processing unit 55, the determination unit 70 determines the appearance of a mouse, and the result of the determination of the appearance is acquired by the imaging unit 20 for the detection of the mouse. Is stored in the storage unit 60 together with the time when the image of the image is captured.

  For example, when the area is within the range of the mouse determination threshold set in advance, the determination unit 70 determines that the mouse is detected as the appearance determination, and when the area is outside the range of the mouse determination threshold. Then, the determination of no mouse detection is performed as the appearance determination.

  The upper limit value of the range of the mouse determination threshold is the case where the imaging unit 20 captures the first reference size mouse with visible light and infrared rays at the nearest position in the mouse detection target area E from the imaging position of the imaging unit 20. This is the size of the area of the image of interest obtained in. In this case, the mouse of the first reference size is preferably the maximum expected size of the mouse.

  The lower limit value of the range of the mouse determination threshold is that the imaging unit 20 captures a mouse having a second reference size from the imaging position of the imaging unit 20 at the most distal position in the mouse detection target area E with visible light or infrared light. This is the size of the area of the image of interest obtained in this case. The mouse of the second reference size in this case is preferably the smallest expected mouse size (eg, a child mouse).

  In addition, as a result of the irradiation light from the projector 40 being reflected on the retina of the pair of eyes of the mouse, a pair of images of interest corresponding to the eye region is obtained. In this case, the center of gravity and the distance between the centers of gravity of the pair of images of interest are calculated by the feature extraction processing unit 55. When the distance between the centroids is within a preset range of the mouse eye distance threshold, the determination unit 70 determines that the mouse is detected as the appearance determination.

  The determination unit 70 combines a plurality of determination methods as described above, and finally determines the determination of the presence of a mouse when there is a determination of a mouse. In addition, the determination unit 70 finally determines the determination that there is no detection of a mouse when the detection of the presence of a mouse is not detected. Then, the determination unit 70 stores the mouse in the storage unit 60 in association with the time (time data) when the image that is the basis is captured in both the determination of whether the mouse is detected and the determination that the mouse is not detected. To do.

Since there are various determinations of whether or not a mouse is detected, it is not limited to the method described above.
The appearance determination of the rat stored in the storage unit 60 and the time when the image that is the basis for the detected mouse detection is captured is output to an output unit (display, printing unit) (not shown) connected to the determination device 30. The

  The determination device 30 includes a communication unit 80 that communicates with an external device. The communication unit 80 may include, for example, a LAN (Local Area Network) interface capable of wired connection or a LAN board capable of wireless LAN connection. Note that the communication method of the communication unit 80 is not limited to the LAN, and may be communication other than the LAN.

(Evaluation apparatus 100)
An evaluation apparatus 100 shown in FIG. 1 is an apparatus that evaluates the mouse detection capability of the mouse detection system 10. The evaluation apparatus 100 includes a communication unit 110 that performs wireless communication with the mouse simulated body 200 and a communication unit 160 that performs communication with the communication unit 80.

  Further, the evaluation device 100 includes a trajectory calculation unit 120, an existence trajectory storage unit 130, a collation unit 140, and an evaluation provision unit 150. The trajectory calculation unit 120 is configured in the area E based on the steering amount of the steering motor M1 of the traveling wheel 230 and the rotation amount of the driving motor M2 of the traveling wheel 230 shown in FIG. The position of is calculated. Note that the initial position of the mouse simulated body 200 in the mouse detection target area E is input as an initial value to the trajectory calculation unit 120 of the evaluation apparatus 100 via an input device (not shown) in advance.

The steering amount and the rotation amount correspond to parameters for calculating the existence locus of the mouse simulated body 200.
In addition, the trajectory calculation unit 120 gives the present trajectory storage unit 130 the current position calculated by the trajectory calculation unit 120 and the time when the mouse is located at the position of the mouse simulation body 200 in the mouse detection target area E. It is memorized as an existence locus. Moreover, the collation part 140 performs the collation process mentioned later. The evaluation assigning unit 150 assigns the evaluation of the determination device 30 based on the collation result of the collation unit.

(Mice mouse 200)
As shown in FIG. 2, the mouse simulation body 200 includes a main body 200a having a mouse simulation shape as a whole, an automatic traveling unit 220 having a plurality of traveling wheels 230 provided at a lower portion of the main body 200a, and an automatic traveling unit 220. Is provided with a control device 240 that controls traveling and steering. Further, as shown in FIG. 2, a reflection material 210 as a pair of reflectors is provided at the front portion of the mouse simulated body 200 so as to be exposed. The pair of reflectors 210 is formed in a murine eyeball. In addition, when providing the reflective material 210, it is preferable that the light projector 40 projects infrared rays. In this case, the light projector 40 corresponds to an infrared light projecting unit. In this case, the imaging unit 20 is preferably an infrared camera.

  Of the plurality of traveling wheels 230 (for example, two front wheels and two rear wheels), the front wheels are steered wheels that are steered by the steering motor M1. The steering wheel is provided with a steering amount sensor A for detecting the steering amount of the steering motor M1. The automatic traveling unit 220 includes a drive motor M2 that rotationally drives the traveling wheel 230 and a rotation amount sensor B that detects the rotation amount of the drive motor M2.

  The control device 240 controls the automatic traveling unit 220 based on a preprogrammed traveling program so that the entire region within the mouse detection target area E (not shown) travels without gaps.

  Further, as shown in FIG. 2, the mouse simulation body 200 includes a communication unit 250 that performs wireless communication with the communication unit 110 of the evaluation apparatus 100. The control device 240 informs the communication unit 110 of the steering amount and rotation amount detected by the steering amount sensor (not shown) and the rotation amount sensor, and the time when these amounts are detected via the communication unit 250. Send it.

(Operation of the embodiment)
Next, the operation of the mouse detection system 10 and the evaluation apparatus 100 of this embodiment will be described.
As shown in FIG. 1, the mouse simulated body 200 is automatically traveled based on the travel program in the entire region of the mouse detection target area E, and the mouse detection system 10 is traveling in the mouse detection target area E by the imaging unit 20. The mouse simulated body 200 is imaged.

  As shown in FIG. 3, in the determination device 30 of the mouse detection system 10, the difference processing unit 53 of the image processing unit 50 performs a difference process between the background image and the one-frame image acquired by the imaging unit 20. Get the difference image. The binarization processing unit 54 binarizes the difference image using a preset threshold value. The feature extraction processing unit 55 labels each pixel of the target image obtained by the binarization process, and performs feature extraction processing for each target image.

  FIG. 5 is an example of a background image of a mouse detection target area E, and FIG. 6 is an example of a one-frame image obtained when a mouse simulated body is present in the mouse detection target area E. FIG. 7 shows an example in which the images of FIGS. 5 and 6 are binarized with a difference image. As shown in FIG. 7, in this example, the feature extraction processing unit 55 performs feature extraction on the part of the reflective material 210 that is the eye of the mouse simulation body 200.

  In the feature extraction processing unit 55, the appearance of the mouse is determined by the determination unit 70 based on the feature extracted for each target image, and the determination result of the appearance is acquired by the imaging unit 20 and related to the target image. The time when an image of one frame is captured is stored in the storage unit 60. In the appearance determination, for example, as shown in FIG. 7, if a pair of images of interest is obtained, the distance between the centers of gravity of the pair of images of interest is the distance between the centers of gravity of the eyes of the mice that are expected to appear. This is performed depending on whether or not it is within the range (the range of the threshold between the centroids). If the distance between the centers of gravity of the image of interest is within the range between the centers of gravity, it is determined that a mouse has appeared.

  In addition, when the whole mouse is imaged, the area in the whole image of the mouse is calculated from the image of interest, and the calculated area is within the range of the area threshold of the mouse that is expected to appear. It is determined that a mouse has appeared.

  On the other hand, when the vehicle automatically travels in the mouse detection target area E, the rotation amount of the drive motor M2 that drives the traveling wheel 230 from the mouse simulated body 200 and the steering amount of the steering motor M1 are evaluated via the communication units 250 and 110. 100 gets. FIG. 1 shows an example in which the mouse simulated body 200 automatically travels as indicated by an arrow.

  In the evaluation apparatus 100, the trajectory calculation unit 120 receives the (acquired) from the mouse simulated body 200 in the area E based on the steering amount of the steering motor M1 of the traveling wheel 230 and the rotation amount of the driving motor M2 of the traveling wheel 230. Calculate the position. The trajectory calculation unit 120 causes the presence trajectory storage unit 130 to store the calculated position and the time when the mouse simulated body 200 is positioned at the calculated position.

  As shown in FIG. 4, the collation unit 140 performs the difference image (accurately, the image of interest) obtained from the image acquired at the same time as the time in order from the initial position of the mouse simulated body 200 in S10. Then, a collation process is performed to determine whether or not it is determined that the mouse is detected. In S <b> 20, the evaluation assigning unit 150 evaluates the mouse detection system 10 based on the collation processing result of the collation unit 140.

  Specifically, when it is determined that there is a mouse detection in the difference images (more precisely, the image of interest) that matches the time of each position of the presence locus of the mouse simulated body 200 at all times, the mouse detection system 10 is evaluated to be 100% detectable. In addition, when the mouse is not detected by the mouse detection system 10 at some time, the imaging unit 20 detects the mouse simulated body 200 in the area of the mouse detection target area E including the existence locus at that time. Evaluate that it is not possible. In FIG. 1, in the mouse detection target area E, the arrow indicated by a dotted line is a case where the direction indicating the movement of the mouse simulated body 200 and its existence locus are made, and in the case where it is determined that the mouse is not detected. Is shown.

In this way, when the evaluation apparatus 100 automatically travels the mouse detection target area E of the mouse simulation body 200, the mouse detection system 10 is evaluated.
In this way, the evaluation program of the evaluation assigning unit 150 evaluates the case where all the existence trajectories can be detected in the trajectory traveled by the mouse simulated body 200 in the mouse detection target area E as being able to guarantee 100%. To do.

According to the evaluation method of the present embodiment described above, the following effects can be obtained.
(1) In the evaluation method of the present embodiment, the appearance determination of the mouse simulated body 200 is performed by the processing of the imaging unit 20, the image processing unit 50, and the determination unit 70 in a state where the mouse simulated body 200 is moved in the mouse detection target area E. Remember. In addition, the presence locus of the mouse simulated body 200 in the mouse detection target area E is stored in the presence locus storage unit 130 separately from the appearance determination by the mouse detection system 10. Moreover, the evaluation of the mouse detection system 10 is given by comparing the appearance determination of the mouse simulated body 200 stored in the storage unit 60 with the existence locus of the mouse simulated body 200. As a result, according to this embodiment, the detection function of the mouse detection system installed in the facility can be evaluated.

  (2) Moreover, in the evaluation method of this embodiment, the reference image which imaged the mouse | mouth detection target area E when the mouse | mouth simulation body 200 does not exist is made into a background image. Then, the image of interest is extracted from the image captured by the imaging unit 20 and the background image by the background difference method. As a result, according to the present embodiment, the target image can be easily extracted by the background subtraction method.

  (3) Further, in the evaluation method of the present embodiment, the mouse simulated body 200 is provided with a mouse eyeball reflector and a mouse simulated shape. As a result, in the mouse detection system 10, it is possible to detect a mouse simulated body by the system in the same manner as a mouse.

  (4) Moreover, in the evaluation method of this embodiment, the mouse simulated body 200 includes the reflector 210 (reflector), and the mouse detection system 10 projects the infrared rays toward the mouse detection target area E. (Infrared projector). The imaging unit 20 is an infrared camera.

As a result, according to the present embodiment, the mouse detection target area E can be imaged by the infrared camera even at night or in a dark room.
(5) The mouse simulation body 200 includes an automatic travel unit 220, a steering amount sensor A that detects the steering amount (parameter) of the steering motor M1 and the rotation amount of the drive motor M2 for calculating the locus of the mouse simulation body 200, A rotation amount sensor B is provided. Then, the locus of the mouse simulated body is calculated based on the parameters. As a result, according to the present embodiment, the existence locus can be automatically acquired only by automatically running the mouse simulated body 200.

In addition, the said embodiment can also be changed and implemented as follows.
In the above-described embodiment, the mouse simulated body 200 travels in the mouse detection target area E by a preprogrammed travel program. However, instead of traveling by the travel program, an autonomous travel program may be used. Good. In this case, the mouse simulated body 200 is provided with a measurement sensor (for example, an ultrasonic sensor) that measures the distance of surrounding obstacles, or a travel route is previously formed in the mouse detection target area E with a marker, and the marker You may make it run autonomously according to.

  In the above embodiment, the mouse simulated body 200 is automatically run, but the mouse simulated body 200 may be arranged in the mouse detection target area E by a person. In this case, the arrangement time and the arrangement position are input to the evaluation apparatus 100 by input means (not shown).

  In the above-described embodiment, the mouse simulation body 200 includes the mouse eyeball reflector 210 (reflector), and the shape of the main body 200a is the mouse simulation shape. In addition, the mouse simulated body 200 may be provided with an infrared light emitter. Further, at least one of a murine eye-spherical reflector 210 (reflector), a murine simulated shape, and an infrared light emitter may be provided. By providing the mouse light emitter, the mouse can be thermally sensed by the infrared camera.

In the case where the mouse simulation body 200 includes an infrared light emitter, the mouse detection system 10 preferably has the imaging unit 20 as an infrared camera.
In the case where the mouse simulation body does not include either the reflector or the infrared light emitter, the imaging unit 20 may be a normal camera. In this case, the system preferably includes a projector that illuminates to a dim level.

  The mouse simulated body 200 may be a stationary body that does not have the automatic traveling unit 220. When the automatic travel unit 220 is provided, it is preferable to travel with a travel program that travels randomly within the detection target area as indicated by the arrows in FIG.

  In addition, when the mouse simulated body 200 is a stationary body, a person randomly places the mouse simulated body at an arbitrary position in the detection target area. And every time the stain simulation body is placed, in order to make the detection by the mouse detection system, the detection is started by the input means of the mouse detection device, and the number of cases where it can be detected is counted The count value is stored. In this case, if the number of tests is the same as the count value that can be detected, 100% guarantee is performed.

  Alternatively, if the murine simulated body is a stationary body, it is assumed that there is an existence locus between positions changed in time series, and if the detected existence locus is different from the existence locus input by a person, it cannot be detected. If they match, the detection guarantee area is set.

  Note that the background subtraction method typically includes a frame subtraction method, an average subtraction method, and the like, but any of them may be used. Here, the frame difference method is a preferable method particularly when there is no change in the background. The average difference method is a method that is preferable when the background pixel value varies with time. The average of a plurality of background images in which no moving object is present is used as a background model, and a foreground image is extracted using this. In the frame difference method, a difference between one frame of a moving image and one frame of the next moving image is taken, and the acquired difference image is binarized.

DESCRIPTION OF SYMBOLS 10 ... Mouse detection system, 20 ... Imaging part, 30 ... Determination apparatus, 40 ... Floodlight,
50 ... an image processing unit, 51 ... a background generation unit, 52 ... a background storage unit,
53... Difference processing unit, 54... Binarization processing unit, 55.
60 ... storage unit, 70 ... determination unit,
DESCRIPTION OF SYMBOLS 100 ... Evaluation apparatus, 120 ... Trajectory calculation part, 130 ... Existence trace memory | storage part,
140 ... collation part, 150 ... evaluation grant part,
200 ... mock mouse, 200a ... main body, 210 ... reflector (reflector),
220 ... automatic running unit, 230 ... running wheel, 240 ... control device,
A ... steering amount sensor, B ... rotation amount sensor, E ... mouse detection target area.

Claims (6)

An imaging unit that captures a mouse detection target area; an image processing unit that extracts a target image from an image captured by the imaging unit; and a feature extraction of the target image; and the appearance of a mouse is determined based on the feature extraction In an evaluation method for a mouse detection system having a determination unit,
In the state where the mouse simulation body is arranged or moved in the mouse detection target area, the appearance determination of the mouse simulation body is stored by the processing of the imaging unit, the image processing unit, and the determination unit, and the mouse A first step of calculating an existence locus of the mouse simulated body in the rat detection target area separately from the appearance determination by a detection system and storing the existence locus in a presence locus storage unit;
A method for evaluating a mouse detection system comprising a second step of comparing the presence determination of a mouse simulated body stored in the presence locus storage unit with the presence locus of the mouse simulated body to give an evaluation of the mouse detection system.   The image processing unit uses a reference image obtained by imaging the mouse detection target area when the mouse simulated body does not exist as a background image, and uses the image obtained by the imaging unit and the reference image as a background difference method to calculate the image of interest. The method for evaluating a mouse detection system according to claim 1, wherein the mouse is extracted.   The mouse detection system evaluation method according to claim 1 or 2, wherein the mouse simulation body includes at least one of a mouse eyeball reflector, a mouse simulation shape, and an infrared light emitter.   The mouse detection system evaluation method according to claim 3, wherein the mouse simulation body includes the infrared light emitter, and the imaging unit is an infrared camera.   The mouse simulation body includes the reflector, and the mouse detection system includes an infrared light projecting unit that projects infrared light toward the mouse detection target area, and the imaging unit is an infrared camera. The evaluation method of the mouse | mouth detection system of description. The mouse simulated body includes an automatic traveling unit that performs automatic traveling, and a sensor that detects a parameter for calculating an existence locus of the mouse simulated body when traveling by the automatic traveling unit,
The mouse detection system evaluation method according to any one of claims 1 to 5, wherein, in the first step, an existence locus of the mouse simulated body is calculated based on the parameter.