JP7261467B2 - animal trap - Google Patents

Description

TECHNICAL FIELD The present invention relates to an animal trapping device capable of trapping only preset animals regardless of day or night.

In recent years, damage to crops by wild birds and animals has become a problem, and measures against wildlife damage have been taken nationwide under the initiative of the Ministry of Agriculture, Forestry and Fisheries. The capture of vermin such as deer and wild boar has become a particular problem.

To trap vermin such as deer and wild boar, cages that close the shutters by means of treads or kicks are often used. However, there is a problem that the vermin cannot be captured unless the vermin enters in such a manner that these physical mechanisms work well. In addition, animals that enter such cages are not limited to vermin such as deer and wild boars, and dogs, cats, and the like may enter, which poses a problem of malfunction. Therefore, a capture device has also been proposed that determines whether an animal is an animal to be captured based on an image captured by an imaging device, and closes the shutter only when the animal is an animal to be captured (for example, Patent Document 1). .

JP 2018-183094 A

There is a problem that it is difficult to automatically determine the type of animal based on the captured image, for example, for animals that have similar appearances, such as dogs and raccoon dogs. In addition, many of the animals to be captured are nocturnal, and in order to distinguish the nocturnal target animals from captured images, it is necessary to adopt a device configuration that enables nighttime imaging.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an animal capturing apparatus capable of determining whether or not an animal is a predetermined animal based on a captured image regardless of day or night, and capturing the animal.

The animal trapping apparatus of the present invention comprises a cage having a shutter that closes an entrance, a locking member that fixes the shutter at a raised position, a shutter driving device that closes the shutter at the lowered position , and an animal in the cage. A first imaging device installed in a cage that captures an external light image, and a control device that analyzes the infrared light image of the first imaging device and determines whether it is a preset animal. , the control device stores a storage unit that stores a first machine-learned model generated using infrared light image data of the animal, and a captured image obtained by the first imaging device; By applying the first trained model, a determination unit that determines whether a preset animal has entered the cage, and the entry of the preset animal is confirmed based on the determination result of the determination unit. a shutter driving unit that causes the shutter driving device to perform a closing operation when the shutter is driven; and an animal detection sensor that detects an animal existing in the imaging range of the first imaging device, wherein the animal detection sensor detects the shutter. characterized by being provided in
The animal trapping device may further include an infrared illumination device for irradiating an imaging range of the first imaging device with infrared light.
In the above animal trapping device, the first imaging device may include an imaging lens fitted with an infrared filter.
In the above animal trapping device , the control device may include a lighting control section that emits infrared light from the infrared lighting device when the animal detection sensor detects an animal.
The animal trapping device further includes a second imaging device that captures a visible light image of the animal in the cage, and the control device includes machine-learned image data generated using visible light image data of the animal. The second trained model is stored in the storage unit, and the determination unit applies the captured image obtained by the second imaging device during the daytime to the second trained model, thereby It is determined whether or not a preset animal has entered the cage, and at night, the captured image obtained by the first imaging device is applied to the first trained model, thereby allowing the preset animal to enter the cage. It may be characterized by determining whether the
The above animal trapping device further comprises a visible light illumination device for irradiating visible light to drive off and/or a sound generator for generating sound to drive off, wherein the image capturing device is positioned before entering the cage. An alarm unit capable of capturing an image of an animal, and operating the visible light illumination device and/or the sound generation device when the determination unit determines that the animal is an animal other than a preset animal. may be characterized.

According to the present invention, it is possible to provide an animal capturing device capable of determining whether or not an animal is a preset animal based on a captured image regardless of day or night, and capturing the animal.

1 is a block diagram of an animal trap according to an example embodiment; FIG. It is a figure explaining (A) an infrared-light image of a wild boar and (B) a visible-light image of a wild boar according to an embodiment. 1 is a perspective view of a cage with an animal trap according to an example embodiment; FIG. 1 is a perspective view of a shutter driving device according to an embodiment; FIG. It is a figure explaining the operation|movement flow of the animal capture apparatus which concerns on an embodiment. (A) Recognition example of a stuffed boar using a visible light image, and (B) Recognition example of a stuffed dog using a visible light image.

An animal trapping device 1 according to an embodiment will be described below with reference to the drawings.
As shown in FIG. 1, the animal trapping device 1 according to the embodiment includes a control device 2, an imaging device 3, a shutter driving device 4, an infrared lighting device 5, an animal detection sensor 6, and a power supply 7. and installed in a cage 10 for catching wild boars.

The control device 2 is a computer including a storage unit 21, a determination unit 22, a shutter driving unit 23, an illumination control unit 24, and a communication unit 25. In this embodiment, the Raspberry Built with Pi.
The storage unit 21 is configured by a storage such as an SSD (Solid State Drive), a microSD card, or the like that stores the control program and the learned model.

The determination unit 22 is a function of a control program that analyzes the infrared light image of the imaging device 3 and determines whether it is a preset animal. In this embodiment, in order to enable high-speed processing of object detection, the determination unit 22 is constructed using Neural Compute Stick 2 (Intel), and as an object detection model for detecting a boar, operation speed A fast MobileNet-based SSD model was adopted. MobileNet is one of the models of CNN (Convolutional Neural Network), which is a deep learning technique, and is released by Google. Of course, it is also possible to adopt a trained model trained by a CNN model other than the illustrated MobileNet. As training data, an independently collected infrared light image of a wild boar was prepared, and trained on an external PC to generate a trained model. FIG. 2A shows an example of an infrared light image of a wild boar, and FIG. 2B shows an example of a visible light image of a wild boar.

The shutter drive unit 23 is a function of a control program that causes the shutter drive device 4 to perform a closing operation when entry of a wild boar, which is a preset animal, is confirmed based on the determination result of the determination unit 22 .
The lighting control unit 24 is a function of a control program that causes the infrared lighting device 5 to emit infrared light when the animal detection sensor 6 detects an animal. When the animal detection sensor 6 does not detect an animal, the infrared light illumination device 5 is not operated, so the battery consumption of the power supply 7 can be reduced. Note that when the animal detection sensor 6 does not detect an animal, the imaging device 3 may not perform imaging.

The communication unit 25 includes a wireless module that enables wide-area wireless communication over several kilometers such as 3G, 4G, and LPWA (Low Power Wide Area). The operator is notified by e-mail or the like that the shutter drive unit 23 has been operated. At this time, an image of the interior of the cage 10 captured by the imaging device 3 may be sent to the administrator. Since the administrator can grasp that the shutter drive unit 23 has been operated by e-mail or the like, it is not necessary to regularly patrol the cage 10. - 特許庁Note that the communication unit 25 is not an essential component, and may be omitted if the device cost needs to be reduced.

In this embodiment, the imaging device 3 is configured by attaching an infrared filter to a lens of a USB camera sensitive to infrared light. An infrared filter is a filter having a function of transmitting only infrared rays of a specific wavelength and shutting out visible rays, and is sometimes called an IR filter. The operating speed of object detection is about 4 FPS for a 1280×780 pixel image obtained from a USB camera, which is sufficient speed for capturing a wild boar. In this embodiment, since the target animal is imaged by the infrared light image not only at nighttime but also during the daytime, it is possible to determine the target animal with the same device configuration both during the daytime and at nighttime. Note that unlike the present embodiment, the imaging device 3 may be configured by an infrared camera, and in this case, a band-pass filter that transmits only infrared light in a specific wavelength band may be combined. In addition, the imaging device 3 is configured to include both a visible light camera and an infrared light camera, and the determination unit 22 has a learned model for determining whether the animal is a preset animal based on a visible light image during the daytime, and an infrared camera during the nighttime. A configuration may be used in combination with a trained model that determines whether the animal is an animal that has been set in advance using an optical image.

FIG. 3 is a perspective view of cage 10 having animal trapping device 1 according to an embodiment.
The cage 10 includes an entrance 11, a shutter 12 for opening and closing the entrance 11, and a shutter support frame 13, and is sometimes called a box trap. A shutter driving device 4 is arranged on the shutter support frame 13 . A bait is placed in the cage 10 to lure the wild boar, which is the target animal.

The shutter driving device 4 includes a motor 41, a locking member 42, a movable member 43, and a rotating member 44, as shown in FIG. A locking member 42 connected to the shutter 12 functions as a holding member that fixes the shutter 12 at the raised position. At this time, the movable member 43 that is reciprocated by the motor 41 is at the first position (for example, the advanced position) and supports the locking member 42 via the rotating member 44 . When the shutter drive unit 23 issues a closing command, the motor 41 moves the movable member 43 to the second position (for example, the retracted position), thereby rotating the rotating member 44 and releasing the support by the locking member 42 . Therefore, the shutter 12 drops to the lowered position and closes the entrance 11 . As a result, the wild boar is captured in the cage 10. - 特許庁
The shutter 12 may be configured to drop by its own weight, but it may be configured to increase the drop speed by providing a biasing mechanism in the drop direction made of an elastic body such as a spring or rubber in the shutter support frame 13 . . The shutter 12 may be provided with a locking mechanism that prevents the shutter 12 from being lifted from the lowered position.

The animal detection sensor 6 is an infrared sensor that periodically senses whether an animal exists within the imaging range of the imaging device 3, and transmits a detection signal to the control device 2 when an animal is detected. When the control device 2 receives the animal presence detection signal from the animal detection sensor 6 , the infrared lighting device 5 irradiates the imaging range of the imaging device 3 with infrared light.
The power supply 7 includes a battery that supplies power to the control device 2 , imaging device 3 , infrared lighting device 5 and animal detection sensor 6 . A solar battery panel may be provided on or near the top surface of the cage 10 to supply power to the power supply 7 .

FIG. 5 is a diagram explaining the operation flow of the animal trapping device 1 according to the embodiment.
When the animal capture device 1 disposes the shutter 12 at the raised position and the monitoring mode is executed, the animal detection sensor 6 performs monitoring (S101). When the animal detection sensor 6 detects an animal (S102), the controller 2 causes the infrared lighting device 5 to emit infrared light (S103). Subsequently, the control device 2 analyzes each frame of the infrared light image captured by the imaging device 3, and determines whether the target animal is a wild boar (S104). The control device 2 counts the number of detected frames each time the target animal (boar) is detected in each frame of the infrared light image (S105), and counts the set number of frames or more (for example, 30 frames or more) within the set time. (S107), the shutter drive device 4 is operated (S108). Here, the reason why the set number of frames is provided is to prevent the shutter driving device 4 from operating when the target animal (wild boar) enters the cage 10 with vigilance and immediately leaves the cage 10 . When the shutter driving device 4 is operated, a notification of shutter operation is sent to the administrator by e-mail or the like (S109).
If the animal captured in the infrared light image captured by the imaging device 3 is a non-target animal such as a dog or cat (No in S104), the irradiation of the infrared light is stopped after a certain period of time has elapsed, Returning to S101 again, the monitoring by the animal detection sensor 6 is continued.

When a stuffed boar is placed on a self-propelled trolley that can be remotely controlled and made to enter the cage 10, the recognition result that it is a wild boar is displayed on the display connected to the control device 2 and is recognized. When the number of frames reached 30 frames, the shutter driving device 4 was operated. When a stuffed animal dog is placed on the same self-propelled cart and entered into the cage 10, the recognition result that it is a dog is displayed on the display connected to the control device 2, and the shutter driving device 4 is activated. didn't work.

Although the preferred embodiments of the present invention have been described above, the technical scope of the present invention is not limited to the description of the above embodiments. Various modifications and improvements can be added to the above-described embodiment examples, and forms with such modifications and improvements are also included in the technical scope of the present invention. For example, wild boars and deer are exemplified as pests, but the present invention is not limited to these, and can be applied not only to monkeys, muntjacs, palm civets, raccoon dogs, raccoons, badgers, etc., but also to pests such as pigeons and crows. be.

In addition, a visible light illumination device for flickering visible light to ward off animals and/or a sound generator for generating sound to drive away animals is additionally provided, and the imaging device 3 is used to ward off the animals before they enter the cage 10 . is imaged, and when the determination unit 22 determines that it is an animal other than a preset animal (for example, in the case of a cat or dog), a visible light illumination device and / or an alarm unit that operates a sound generation device may be provided in the control device 2. According to such a configuration, it is possible to prevent the food in the cage 10 from being eaten by animals other than the target animal.

Also, an automatic feeder may be installed in the cage 10 and the control device 2 may monitor the remaining amount of food based on the image captured by the imaging device 3 . Further, it may be configured such that food is automatically replenished after a certain period of time has elapsed, or the automatic feeder may be configured to be remotely controlled by wide area wireless communication. Wild boars are sensitive to the smell of bait, and the catch rate can be increased by always setting fresh bait. As for the automatic feeder, commercially available ones for dogs and cats can be used as they are or after being improved.

Further, an ultraviolet light illumination device is adopted in place of the infrared light illumination device 5, the imaging device 3 is configured by an ultraviolet light camera, and the determination unit 22 is trained to determine whether or not it is an animal set in advance by ultraviolet light. It may be configured by a model. Furthermore, the imaging device 3 is configured to include both a visible light camera and an ultraviolet light camera, and the determination unit 22 has a learned model that determines whether the animal is a preset animal based on a visible light image during the daytime and an ultraviolet light image during the nighttime. A configuration may be used in combination with a trained model that determines whether the image is an animal set in advance. FIG. 6 shows (A) a recognition example of a stuffed boar using a visible light image, and (B) a recognition example of a stuffed dog using a visible light image.

The shutter driving device 4 is not limited to the illustrated configuration, either. It is good also as a structure provided with the motor which makes it advance/retreat.

REFERENCE SIGNS LIST 1 animal trapping device 2 control device 3 imaging device 4 shutter drive device 5 infrared lighting device 6 animal detection sensor 7 power source 10 cage 11 entrance 12 shutter 13 shutter support frame

Claims (6)

A cage having a shutter that closes the entrance;
a locking member that fixes the shutter in the raised position;
a shutter driving device that closes the shutter to the lowered position ;
a first imaging device installed in the cage that captures an infrared light image of the caged animal;
A control device that analyzes the infrared light image of the first imaging device and determines whether it is a preset animal,
a storage unit in which the control device stores a machine-learned first learned model generated using infrared light image data of the animal;
a determination unit that determines whether an animal preset in a cage has entered a cage by applying a captured image obtained by the first imaging device to the first trained model;
a shutter driving unit that causes the shutter driving device to perform a closing operation when an animal is confirmed to enter, which is preset based on the determination result of the determining unit ;
Furthermore, an animal detection sensor that detects an animal existing in the imaging range of the first imaging device,
The animal trapping device , wherein the animal detection sensor is provided on the shutter . 2. The animal trapping apparatus according to claim 1, further comprising an infrared illumination device for irradiating an imaging range of said first imaging device with infrared light. 3. The animal capture device of claim 2, wherein said first imaging device comprises an imaging lens fitted with an infrared filter. 4. The animal trapping device according to claim 2, wherein the control device comprises a lighting control section that emits infrared light from the infrared lighting device when the animal detection sensor detects an animal. . Furthermore, comprising a second imaging device that captures a visible light image of the caged animal,
wherein the control device stores a machine-learned second learned model generated using the visible light image data of the animal in a storage unit;
The determination unit determines whether an animal preset in the cage has entered the cage by applying the captured image obtained by the second imaging device to the second trained model during the daytime, and 5. Whether or not a preset animal has entered a cage is determined by applying the captured image obtained by said first imaging device to said first trained model. The animal trapping device according to any one of 1. Furthermore, a visible light illumination device for irradiating visible light to ward off and / or a sound generator for generating sound and warding off,
The imaging device can image the animal before entering the cage,
1. An alarm unit for operating the visible light illumination device and/or the sound generator when the determination unit determines that the animal is an animal other than a preset animal. 6. The animal trapping device according to any one of 5.

Images