JP2017194760A - Method and system for investigating inhabitation status of animal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly measure the number of specific animals in each of predetermined areas.SOLUTION: An animal habitation status investigation method includes: an imaging step of a capturing image with an infrared thermography camera mounted on a flying object, which flies in a preset flight route, when arriving at an imaging position or periodically, and storing the image with an imaging position including latitude, longitude, and altitude on storage means; an image analyzing step of detecting a specific animal from the captured image stored in the storage means, in a computer device; and a statistical processing step of estimating the number of animals in all areas, by use of the number of animals in a certain area extracted by the image analyzing step.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an animal habitat survey method and an animal habitat survey system that investigate an animal habitat using a flying object such as a drone.

  Currently, when grasping the number of birds and beasts that live in a certain area or range, we first visually measure a limited narrow range, and based on this data, predict the number of individuals living in a certain range. . In such a visual method, the number of individuals cannot be accurately grasped, and the observation range becomes narrow, so it is difficult to accurately determine the effect when the number of individuals is adjusted by capture / hunting. .

  On the other hand, techniques for identifying birds and beasts in images based on images of birds and animals have been proposed. For example, in Japanese Patent Laid-Open No. 2004-228867, a pre-created stationary bird form is specified by performing an erasing process for erasing moving objects from video data obtained by photographing a stationary bird for a predetermined time. A method has been proposed in which template matching is performed between the reference data and the target data to determine the number of stationary birds and the stationary position.

  In Patent Document 2, inter-frame difference processing is performed on a plurality of images acquired by photographing a specific spatial region, while the frame is divided into a plurality of small regions, and the region is further divided into a birds and beasts body. Is divided into a fuselage area assumed as a wing part of a bird and beast, and the small area is fluttered on the basis of the size of the number of pixels obtained by difference processing between the fuselage area and the wing area. There has been proposed a technique for discriminating a fluttering state of a wild beast among the wild animals existing in a specific spatial region by determining whether or not the present wild bird is an area.

JP 2011-138310 A JP 2014-92842 A

  However, although the techniques of Patent Documents 1 and 2 can be applied to birds, it is difficult to apply them to animals such as deer and frogs.

  The present invention has been made in view of the above-described circumstances, and can be applied not only to birds but also to animals that move around in a wide range, and can accurately measure the number of individuals of a specific animal for each fixed area. An object is to provide an animal habitat survey method and an animal habitat survey system.

In order to achieve the above object, the animal habitat survey method according to the present invention is:
A flying object equipped with an infrared thermography camera is made to fly along a preset flight route, and when the shooting position is reached, the flying object is stopped in the air and shot with an infrared thermography camera. A shooting stage that stores the shooting position and time including altitude in a storage means, and
An image analysis step of inputting data stored in the storage means to a computer device and detecting a specific animal from the captured image;
The mesh data is created using a Markov chain Monte Carlo method or the like, and the mesh data and the number of animals extracted in the image analysis stage are used to include the animal in the entire target area including the imaging range. And a statistical processing stage for estimating the number of individuals.

In addition, the animal habitat survey method according to the present invention includes:
A GPS, an infrared thermography photographing means, and a storage means for storing a photographed image photographed by the infrared thermography camera;
A computer apparatus for detecting a specific animal from a photographed image stored in the storage means, and an animal habitat survey system comprising:
The aircraft is
Flight control means for flying based on a preset flight route or a command from a remote control, detecting that the shooting position has been reached, and stopping the flying object in the air;
A photographing control means for photographing the lower part with an infrared thermography camera, attaching the photographing position information including the latitude, longitude, and altitude to the photographed image and the time, and storing it in the storage means;
The computer device includes:
Image analysis means for detecting a specific animal from the captured image stored in the storage means;
Statistical processing means for creating mesh data using a Markov chain Monte Carlo method or the like, and estimating and calculating the number of animals in the entire target area using the mesh data and the number of animals detected by the image analysis means When,
Means for outputting a result of estimation calculation by the statistical processing means;
It is provided with.

  In particular, the image analysis means of the computer device extracts a predetermined pattern shape from the luminance image of the captured image, and detects a target animal based on the temperature distribution of the pattern shape. Since the animal has a different temperature distribution for each type, the individual position is specified from the luminance image, and it is determined whether the animal is a detection target animal based on the temperature distribution of the individual position in the thermo image.

Then, the statistical processing means inputs the number of individuals in the mesh to mesh data obtained by meshing the target area, thereby estimating and calculating the number of individuals in the entire target area.
Thereby, a specific animal can be detected with high accuracy, and the number of individuals in the entire area can be obtained quickly and easily. In particular, when a drone is used as a flying object, the drone has a great effect because the flightable time is limited.

  As described above, according to the present invention, it is possible to accurately measure the number of individuals of a specific animal for each certain area at a lower cost than in the past.

It is a functional block diagram of the animal habitat situation investigation system by embodiment of this invention. It is a data block diagram of the thermo image data file in the memory | storage means of FIG. It is a flowchart which shows the process sequence of the flight control means of FIG. It is a flowchart which shows the process sequence of the imaging | photography control means of FIG. It is a flowchart which shows the process sequence of the image analysis means of FIG. It is explanatory drawing of the data structural example of pattern DB of FIG. It is a figure which shows the mesh data example displayed on a computer screen by the output means of FIG.

  Hereinafter, a first embodiment of an animal habitat survey system according to the present invention will be described with reference to the drawings.

  Referring to FIG. 1, an animal habitat survey system 1 according to the present embodiment schematically includes an infrared thermography camera (hereinafter referred to as “thermo camera”) 22, a drone (aircraft) 10 equipped with a GPS function, The computer device 30 is configured to detect a specific animal based on an image photographed by the camera 22 (hereinafter referred to as a “thermo image”) and to perform a calculation process of the number of individuals of the animal.

<Preparation stage>
The flight route of the drone 10 is performed by selecting a plurality of points on the map. The selected point is stored in a memory included in the drone as latitude and longitude information. In addition to the flight route information, altitude information can be input, and the direction magnification of the camera (visible light camera) 21 included in the drone may be set.

  The drone 10 also includes a thermo camera 22. The user sets in advance the shooting position (latitude, longitude, altitude) of the thermo camera 22. The photographing position of the thermo camera 22 is set so that the photographing ranges overlap to some extent from the magnification and altitude. Instead of setting the shooting position in advance, the shooting may be performed at a constant cycle based on the moving speed of the flying object 10. This setting information is stored in the storage means 13 included in the drone 10.

<Photographing stage>
The drone according to the present embodiment is based on autonomous flight along the flight route set by the above-described procedure. However, while the operator confirms the image sent from the visible light camera 21 of the drone 10 ( It is also possible to give a command using a remote control. In this case, the command sent from the controller is prioritized, and the drone 10 flies based on the command.

  Thus, the drone 10 flies based on the command from the controller by the flight control means 11 or autonomously flies along the flight route inputted in advance. The shooting control means 12 of the drone 10 always collects the position information of the drone 10 from the GPS, and when the drone 10 reaches a preset position, it releases the shutter of the thermo camera 22 and acquires image data. Note that the imaging control means 12 may periodically release the shutter of the thermo camera 22. At this time, the period may be made variable according to the flying speed of the drone 10 so that the acquired images overlap by a certain width. In addition to this, it is preferable that the shutter of the thermo camera 22 is operated in accordance with an instruction from the operator sent via the controller.

  The image data acquired by the thermo camera 22 is stored in a removable storage means 13 (for example, an SD card) provided in the drone 10. FIG. 2 shows a data configuration example of the thermo image data file 13a stored in the storage means 13. Shooting conditions such as a shooting position, a shooting time, and a magnification are stored in association with each shot thermo image (shot image).

FIG. 3 shows an example of the processing procedure of the flight control means 11.
When the flight control unit 11 is activated, it acquires position information from a GPS or altitude sensor (S101). When it arrives at the photographing position, it stops in the air (S102, S103), activates the photographing control means 12 and acquires a thermo image (S104). Then, it is determined whether or not all photographing has been completed (S105). If not, the next photographing position is set as a target point for flight (S106), and the processing from step S101 is repeated. On the other hand, when all the shootings are completed in step S105, the target point is set as the departure point and the return trip is reached.

  FIG. 4 shows an example of the processing procedure of the imaging control means 12. When activated by the flight control unit 11, the imaging control unit 12 adjusts the altitude of the drone 10 and the magnification of the thermo camera 22 to preset values, acquires a thermo image from the thermo camera 22, and stores the storage unit 13. Save to. In some cases, the drone 10 cannot be moved to a predetermined altitude due to geographical conditions or the like. For this reason, the imaging control means 12 may adjust the magnification in step S204 so that the previously acquired image overlaps the area in a certain range.

  By making the autonomous flight according to the above-described procedure and setting the shooting point so that certain regions of the shot image overlap in advance, the thermo image can be reliably acquired. As a simpler method, the pilot sends a sequential flight command from the pilot to the drone 10, while the imaging control means 12 periodically or every time the camera is moved from the previously captured position by a certain distance from the thermo camera 22. There is a method of acquiring a thermo image.

<Image analysis stage>
When the drone 10 returns from the shooting flight, the user attaches the storage means 13 to the computer device 30 so that the data can be read. The computer device 30 can be realized by a general-purpose computer. Instead of the detachable storage unit 13, the drone 10 may send the data stored in the storage unit 13 to the computer device 30 by wireless communication or the like.

  Then, the image analysis means 31 of the computer device 30 is activated to read the thermo image data stored in the storage means 13, and a specific animal is detected from this image data by the following procedure.

(Example of specific animal detection procedure)
(S1) The individual position is extracted by pattern matching based on the luminance image in the thermo image. At this time, the image analysis means 31 extracts the pattern of the animal to be detected from the pattern DB 41 based on the time of photographing and the ground surface temperature (or air temperature). This pattern includes a temperature distribution in or around the pattern shape in addition to the pattern shape. Since the amount of hair on the skin surface differs depending on the season, for example, the season, and the thermographic image changes depending on the ground surface temperature, the pattern DB 41 includes a detection target for each season and ground surface temperature as shown in FIG. It is preferable to preserve the animal pattern.
(S2) It is further determined whether or not the temperature (and / or temperature distribution) at a predetermined position in the pattern matches a preset temperature (and / or temperature distribution) with respect to the individual position.
(S3) If the results of the determination match, the extracted individual is counted up as the target animal.
(S4) The individual position information extracted based on the GPS information associated with the thermo image is calculated and stored in the storage means of the computer device.

FIG. 5 shows an example of the processing procedure of the image analysis means 31.
When the image analysis means 31 is activated via an input interface (not shown) of the computer device 30, the image analysis means 31 first generates a luminance image of the thermo image (S301) and matches the pattern shape (template) of the pattern DB 41. All the positions of things are extracted (S302). At this time, a pattern corresponding to the photographing time and the ground surface temperature (or substitute with the atmospheric temperature) is extracted. Then, for a certain individual position, the temperature (temperature distribution) at that position is extracted and compared with a pattern (temperature or temperature distribution) that is a set value stored in the pattern DB 41 (S304).
As a result, if the error from the set value of the detection target is within a certain range, the extracted pattern shape is counted up as the detection target (S306).
The above processing is executed for all the individual positions extracted in step S302 (S307, S308), and the count value is stored in the storage means (S309).

<Individual identification stage>
At this stage, an ID is assigned to each individual and the movement is tracked between a plurality of thermo images. For example, if the temperature distribution in the pattern shape and the pattern shape is the same between the two images in a certain range including the overlapping region of the thermo image by the individual identification means 32, that is, if the difference in the comparison amount is equal to or less than the threshold value, Recognize that there is, and attach the same ID. At this time, for a pattern in one image, a search range in another image may be determined from the normal moving speed of the animal.

  With the above processing, the same individual that is imaged redundantly among a plurality of thermo images is excluded, and an accurate number of individuals is obtained for the entire imaging area.

<Statistical processing stage>
The statistical processing means 33 creates mesh data using the Markov chain Monte Carlo method based on the individual extracted in the image analysis stage and the individual identification stage and the GPS information of the image. The number of individuals in the area can be predicted without observing the entire area to be investigated by using the statistical processing means by putting the number of individuals extracted in the mesh data.

<Result output stage>
Based on the mesh data created in the statistical processing step, the output means 34 displays information such as the extraction position and the number of individuals for each area on the user interface of the computer device. FIG. 7 is a display example of mesh data on the user interface. Statistical data for each mesh obtained by the statistical processing means 33 is displayed superimposed on the map. For example, the number of animals to be investigated can be displayed.

  As described above, according to the present embodiment, the number of individuals of a specific animal in the imaging area can be measured with high accuracy and at a lower cost than in the past. As a result, the number of individuals in the entire survey target area including the imaging area can be estimated with high accuracy.

  The present invention is not limited to the above-described embodiment, and can be implemented with various modifications without departing from the spirit of the present invention. For example, a moving image may be taken with a drone, and tracking of an individual, particularly recognition of the same individual in a captured image overlapping region, may be performed while visually checking the image data. At this time, a more accurate determination is possible by combining with the determination result of the individual identification means 32.

1 Animal habitat survey system 10 Aircraft (drone)
DESCRIPTION OF SYMBOLS 11 Flight control means 12 Imaging | photography control means 13 Storage means 13a Thermo image data file 21 Visible light camera 22 Thermo camera 30 Computer apparatus 31 Image analysis means 32 Individual identification means 33 Statistical processing means 34 Output means

Claims (4)

An aircraft equipped with an infrared thermography camera is made to fly along a predetermined flight route, and when it reaches a predetermined shooting position or periodically, it is shot with the infrared thermography camera. A shooting stage with a shooting position including
In the computer device, an image analysis step of detecting a specific animal from the captured image stored in the storage means;
A statistical processing step for estimating and calculating the number of animals in the entire target area based on the number of animals in a certain range extracted in the image analysis step;
An animal habitat survey method characterized by including A flying object comprising GPS, an infrared thermography photographing means, and a storage means for storing a photographed image photographed by the infrared thermography camera;
A computer apparatus for detecting a specific animal from a photographed image stored in the storage means, and an animal habitat survey system comprising:
The aircraft is
Flight control means for flying based on a preset flight route or a command from a controller;
When the predetermined shooting position is reached or periodically, shooting is performed with the infrared thermography camera, and shooting position information including latitude, longitude, and altitude is added to the shot image and stored in the storage unit; and With
The computer device includes:
Image analysis means for detecting a specific animal from the captured image stored in the storage means;
Statistical processing means for estimating and calculating the number of animals in the entire target area based on the number of animals in a certain range detected by the image analysis means;
Means for outputting a result of estimation calculation by the statistical processing means;
An animal habitat survey system characterized by comprising:   The animal inhabiting state according to claim 2, wherein the image analysis means extracts a predetermined pattern shape from the luminance image of the photographed image, and detects a target animal based on a temperature distribution of the pattern shape. Survey system.   A pattern database that stores patterns of animals to be detected for each time and ground surface temperature range is provided, and the corresponding pattern is extracted from the pattern database based on the time of photographing and the ground surface temperature. The animal habitat survey system according to claim 3, wherein an animal is detected.