WO2019186982A1

WO2019186982A1 - Animal habitation investigation method, animal habitation investigation system, and animal habitation investigation device

Info

Publication number: WO2019186982A1
Application number: PCT/JP2018/013495
Authority: WO
Inventors: 政聡佐々木
Original assignee: 株式会社スカイシーカー
Priority date: 2018-03-29
Filing date: 2018-03-29
Publication date: 2019-10-03
Also published as: JPWO2019186982A1; JP7161522B2

Abstract

[Problem] To provide an animal habitation investigation method by which the number of individual animals can be more accurately counted. [Solution] This animal habitation investigation method includes: an imaging step for imaging an area to be imaged by causing a flying body to fly along a set flight route; an extraction step for extracting a still image at a prescribed timing from the captured image; a registration step for registering detection target animal information in a storage means; a detection step for detecting a detection target animal image from the still image on the basis of the registered detection target animal information; a correction step for accepting a correction that, if an image that should not be detected as a detection target animal image was detected, excludes the image from the detection target animal image; a counting step for counting the number of detection target animal images on the basis of the results of the detection step and the results of the correction step; and an output step for outputting the counting results. In the registration step, the detection target animal information can be updated on the basis of the results of the correction step.

Description

Animal habitat survey method, animal habitat survey system, and animal habitat survey device

The present invention relates to an animal habitat survey method, an animal habitat survey system, and an animal habitat survey apparatus.

Patent Document 1 discloses a method for investigating animal habits using an infrared thermography camera.

JP 2017-194760 A

However, the conventional animal habitat survey methods have problems in that they cannot recognize the animal that is the subject of the survey or misrecognize the animal that is not the subject of the survey.

An object of the present invention is to provide an animal habitat survey method that can more accurately count the number of animals.

In the animal habitat survey method of the present invention, a flying object including an imaging unit is caused to fly along a predetermined flight route, and an imaging process of imaging an imaging target area by the imaging unit, and a captured image captured by the imaging unit. Based on the extraction step of extracting a still image at a predetermined timing from, a registration step of registering detection target animal information capable of specifying an animal image to be detected in a storage means, and the registered detection target animal information A detection step of detecting a detection target animal image from the still image, and when the image that should not be detected as the detection target animal image is detected, the image is excluded from the detection target animal image Based on the correction step for accepting correction, the result of the detection step and the result of the correction step, the animal image of the detection target in the imaging target area And the output step of outputting the counting result, and the registration step can update the detection target animal information based on the result of the correction step. It is characterized by.

This makes it possible to more accurately count the number of animals.

In the animal habitat survey method of the present invention, a flying object including an imaging unit is caused to fly along a predetermined flight route, and an imaging process of imaging an imaging target area by the imaging unit, and a captured image captured by the imaging unit. Based on the extraction step of extracting a still image at a predetermined timing from, a registration step of registering detection target animal information capable of specifying an animal image to be detected in a storage means, and the registered detection target animal information A detection step of detecting a detection target animal image from the still image, and when the image to be detected as the detection target animal image is not detected, a correction is made to include the image in the detection target animal image. Based on the correction process to be accepted, the result of the detection process, and the result of the correction process, the number of animal images to be detected in the imaging target area is calculated. And the output step of outputting the result of the counting, and the registration step can update the detection target animal information based on the result of the correction step. And

This makes it possible to more accurately count the number of animals.

Preferably, in the photographing step, the flight route can be set to the flight route and altitude so that the angle of view and the magnification at which the imaging target area is photographed by the imaging means are constant. It is characterized by that.

This makes the shooting magnification and angle constant, so that it is possible to prevent the standard size of the animal that is the subject from being blurred.

Preferably, the predetermined timing can be set by a predetermined time interval or a predetermined frame interval.

This makes it easier to check still images.

Preferably, when the correction process is not performed for a predetermined period, a screen for prompting correction is displayed.

This makes it possible to prompt the operator to make corrections.

The animal habitat investigation system of the present invention includes an imaging unit that images an imaging target area, and an animal that is an object to be investigated based on a flying object that flies along a preset flight route and an image captured by the imaging unit. An animal habitat survey system that outputs the habitat survey results of the animal, wherein the animal habitat survey apparatus extracts a still image at a predetermined timing from the captured image captured by the imaging means Means for registering detection target animal information capable of specifying an animal image to be detected in a storage means, and detecting a detection target animal image from the still image based on the registered detection target animal information And when the image that should not be detected as the detection target animal image is detected, the image is detected as the detection target animal image. Correction means for accepting correction to exclude from, a counting means for counting the number of animal images to be detected in the imaging target area based on a result of the detection means and a result of the correction means, and Output means for outputting a result, wherein the registration means can register the detection target animal information based on the result of the correction means.

This makes it possible to more accurately count the number of animals.

The animal habitat investigation apparatus according to the present invention includes an extraction unit that extracts a still image from a captured image captured by an imaging unit that captures an imaging target area at a predetermined timing, and a detection target animal that can identify an animal image to be detected. Registration means for registering information in storage means, detection means for detecting a detection target animal image from the still image based on the registered detection target animal information, and the detection target animal image to be detected as the detection target animal image When an image that is not an image is detected, a correction unit that accepts correction to exclude the image from the detection target animal image, and a result of the detection unit and a result of the correction unit A counting unit that counts the number of animal images to be detected; and an output unit that outputs a result of the counting. It is possible to register the detection target animal information based on the stage of the result, characterized in that.

This makes it possible to more accurately count the number of animals.

According to the present invention, it is possible to more accurately count the number of animals.

It is a block diagram of an animal habitat investigation system. It is a figure which shows the data structure of the moving image data image | photographed with the camera. It is a schematic diagram of the flight route of a flying object. It is the schematic of the moving image imaging | photography method of a flying body. It is explanatory drawing which shows the angle of view of the camera of a flying body. It is a flowchart which shows the flight control flow of a flying body. It is a flowchart which shows the imaging | photography control flow of a camera. It is a flowchart which shows an animal habitat investigation result processing flow.

Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS.

<Configuration of animal habitat survey system>
FIG. 1 is a configuration diagram of an animal habitat survey system according to this embodiment of the present invention.
In the animal habitat survey system 1 according to the present embodiment, the flying object 2 and the server 3 are communicably connected via a wide area communication network 4 (typically the Internet), and data is exchanged.

[Configuration of the flying object]
First, the configuration of the flying object 2 will be described. The flying object 2 is an unmanned flying object such as a drone. The flying object 2 includes a control unit (flight control unit 21 and imaging control unit 23) for controlling the flying object 2, a driving unit for driving a propeller, a GPS for obtaining position information, a sensor, a camera 22, a RAM and a ROM. The storage means 24 which consists of storage media, such as, a battery, etc. are provided. The control unit controls the camera 22, the sensor, and the GPS to acquire moving image data, still image data, weather data such as temperature and humidity, and position data, and transmits the acquired various data to the server 3 through the communication unit. .

The camera 22 is an imaging means provided in the flying object 2 for confirming shooting for setting and confirming the flight range in the previous stage of conducting the animal habitat survey and for animal habitat survey in the set flight range. The shooting control unit 23 controls moving image shooting and still image shooting.

The flight control means 21 controls the flight of the flying object 2 according to the flight route based on the animal habitat survey flight route data set by the server 3. Specifically, according to the animal habitat survey flight route data (flight route, height, speed, camera angle, magnification, shooting start point, shooting end point, etc.) preset in the server 3 Control the flight. The sensor is a sensor for acquiring weather data such as temperature and humidity. The GPS is a GPS receiver, and is used to fly along a flight route according to a flight control signal transmitted from the server 3. The flight control means 21 acquires flight position information of the flying object 2 by GPS. For example, information on the current position of the flying object 2, information on the position of the flying object according to the flight route transmitted from the server 3, and the like are acquired. The drive unit controls the drive of the propeller according to the control signal transmitted from the server 3. The storage unit 24 stores moving image data and the like taken by the camera 22 along the flight route in association with date / time data, latitude / longitude / altitude data, weather data, and the like.

Note that the flight route of the flying object 2 at the time of the animal habitat survey includes the flight start point and the flight end point of the flight route of the flight object 2 in the animal habitat survey target area (in this embodiment, the flight end point and the flight route). The flight start point is the same), and each transit point from the start point to the end point is overlaid to be digitized (indicated by latitude and longitude), and flight route data is created by setting the flight altitude. In addition, a shooting start point and a shooting end point at each waypoint from the flight start point to the flight end point are set. Then, after setting the flying speed of the flying object 2 and the angle and magnification of the camera 22, the flying object 2 is made to fly based on the set flight route data, and a captured image by the camera 22 is confirmed. Then, after confirming the captured image (captured video), by repeating the work to correct the flight route, the flight altitude, the flight speed, the angle of view and magnification of the camera 22, etc. The flight route data, the flight speed, the angle of view and magnification of the camera 22 are determined. Note that the angle of view of the camera 22 is preferably set to an angle (for example, 45 degrees) at which the characteristics of the animal under investigation can be easily recognized, and the magnification of the camera 22 is set to a constant magnification according to the size of the animal under investigation. Is preferred. The determined flight route data, flight speed, angle of view and magnification of the camera 22 are registered in the DB 32 as animal habitat survey flight route data in association with the animal habitat survey ID, such as “animal habitat survey No. 1”. .

Then, in association with the registered animal habitat survey flight route data, still image data extraction (capture) timing, for example, extraction at intervals of 2 seconds, or at 60 frames / second, 120 frame intervals, etc. are set. The extraction timing of the still image data here is preferably in a state of being output like a single aerial photograph when the extracted still image data is continuously connected and output to a monitor or the like. As for the extraction timing of still image data, a method of batch setting, a method of setting for each still image data, a method of finely adjusting the extraction timing while checking each still image data after batch setting, A method of extracting at an arbitrary timing while checking the above is conceivable.

[Server configuration]
Next, the configuration of the server 3 will be described. The server 3 is an animal habitat investigation device, and includes a control unit (a registration unit 32, an extraction unit 35, an image analysis unit 36, a correction unit 37, and a counting unit 38) including a CPU as a hardware configuration. The control unit includes a communication unit for communicating with an external wired or wireless wide area communication network (typically the Internet), a display unit including a display such as a liquid crystal, and a storage medium such as a ROM, RAM, and HDD. A storage unit is connected. The storage unit stores a database (DB) 31 that holds processing results such as image data and image analysis, an application program for animal habitat investigation, a flight route of the flying object 2, weather data, and the like. The server 3 is connected to input means 33 such as a keyboard and a mouse, and output means 34 such as a display and a printer.

The input means 33 is, for example, a keyboard, mouse, touch panel or the like operated by an operator, and the operator performs various operations of the server 3, a flight route registration operation, a camera 22 pan / tilt registration operation, a still image, etc. Various operations such as an operation for registering individual image information from data and a success / failure operation for detected individual image data are performed.

The output means 34 may be any device that can display a moving image such as a monitor, a still image, or the like. In addition, the apparatus (touch panel display monitor) which has the structure by which the input means 33 and output means 34, such as a touchscreen, were united may be sufficient.

In the DB 31, moving image data of an aerial image captured by the camera 22 is stored in association with an ID, the still image data extracted by the extraction unit 35, and an individual subject to be registered in advance (also referred to as a counting target). The image information, the individual image information to be counted added by the registration means 32, the data necessary for animal habitat investigation including the set flight route and weather data are stored. In the present embodiment, as moving image data, color moving image data captured by the camera 22 is stored in association with date / time data, latitude / longitude / altitude data, weather data, and the like. In addition, the individual image information to be counted includes image information obtained by photographing the whole image of the individual from a plurality of angles, image information obtained by photographing the individual at a wide angle or a telephoto, and individual feature points (for example, eyes, nose, mouth, ears). , The shape of a corner, etc.) from various angles, and various pieces of image information such as a male individual, a female individual, and a child individual. In addition, the individual image information to be counted is the deer / individual image information, the monkey individual image information, the monkey, when there are a plurality of types of animals to be inhabited this time, for example, the deer / moth / monkey. Is stored in the DB 31. Further, when the first correction process by the correction unit 37 to be described later is not executed in the DB 31 for a predetermined period, a timing for displaying on the monitor a first correction process promotion display for prompting the execution of the first correction process (for example, 30 days) ), And when the second correction process by the correction unit 37 is not executed for a predetermined period, the timing (for example, 30 days) for displaying the second correction process promotion display for prompting the execution of the second correction process is registered. Is done.

The registration unit 32 registers a plurality of individual image information (base individual image information) to be counted in advance in the DB 31, or when correction processing is performed by the correction unit 37 described later, the counting target stored in the DB 31. The individual image information is deleted or newly registered as individual image information to be counted.

The extracting unit 35 converts the color moving image data captured by the camera 22 into black and white moving image data, and then extracts black and white still image data at a predetermined timing (predetermined time (several seconds) or every frame). (Capture), and the extracted monochrome still image data is stored in the DB 31 in association with the still image data ID, date / time data, weather data, and the like.

The image analysis unit 36 refers to the individual image information to be counted stored in the DB 31 and performs pattern matching with each still image data (aerial photographed landscape image) extracted by the extraction unit 35. If there is a plurality of pieces of image data having a high matching degree (similarity) with the individual image information to be counted (for example, the matching degree is 85% or more) in the data, this image data is specified as the individual image data to be counted. Then, an image analysis process is performed in which the specific individual image data ID, the still image data ID, the coordinate information, and the like are stored in the DB 31 in association with the specific individual image data. Therefore, when one individual image data to be counted is specified in one still image data, one specific individual image data is stored in the DB 31, and when a plurality of specific individual image data is specified, a plurality of specific individual image data are stored. The Rukoto. Further, the matching degree is stored in the DB 31 in association with each count target individual image information stored in the DB 31.

The correcting unit 37 corrects an error in the specific individual image data specified by the image analyzing unit 36 or adds image data that cannot be specified as the individual image data to be counted by the image analyzing unit 36. Used for. The correction unit 37 is operated by the operator using the input unit 33. When the operator wants to correct an error in the specific individual image data, the input means 33 is operated to display the specific individual image data on the monitor one by one for confirmation. If image data different from the image data of the target animal is included, it can be excluded from the specific individual image data by operating the input means 33 (hereinafter referred to as first correction processing). When the operator wants to check whether there is any image data that cannot be specified as the individual image data to be counted by the image analysis means 36, the input means 33 is operated to input the still image data one by one. Display on the monitor (the still image displayed at this time shows a mark indicating the position where the specific individual image data exists and the type of animal etc. are displayed together with the mark). If there is image data that should originally be specified as specific individual image data, it can be included in the specific individual image data by operating the input means 33 to designate this image data ( Hereinafter, it is referred to as a second correction process).

In addition, when the first correction process described above is performed by the correction unit 37, the registration unit 32 shows individual image information that shows a high degree of matching (for example, 85% or more) with respect to the image data excluded from the specific individual image data. Is deleted from the individual image information to be counted. Note that, when deleting, there are a method in which the operator confirms the individual image information to be deleted and then deletes it according to the operation by the operator, or a method of deleting without confirming by the operator.

In addition, when the above-described second correction process is performed by the correction unit 37, the registration unit 32 performs a process of additionally registering image data newly included in the specific individual image data as individual image information to be counted. In addition, for the additional registration, there are a method in which the operator confirms the image data to be added and then performs additional registration according to the operation by the operator, or a method of additional registration without performing confirmation by the operator.

The counting means 38 determines the number of detected specific individual image data for each specific individual image data corresponding to the type of each animal based on the image analysis processing result by the image analysis means 36 and the correction processing result by the correction means 37 described above. To count.

The counting result by the counting means 38 is output to the monitor or the like by the output means 34. As the output form here, for example, the name of each survey target animal and the number of heads counted are displayed along with the animal habitat date and time.

Here, the animal habitat survey method according to this embodiment will be briefly summarized. In the animal habitat investigation method according to the present embodiment, still image data is extracted from the moving image data photographed by the flying object 2 at a predetermined timing by the extraction means 35, and image analysis processing is performed by the image analysis means 36 from the extracted still image data. The image data having a high matching degree with the individual image information to be counted is stored in the DB 31 as the specific individual image data. When there is an error in the specific individual image data stored in the DB 31, it can be excluded from the specific individual image data by the first correction process by the correction unit 37. Further, the still image data stored in the DB 31 can be confirmed, and when there is image data that has not been detected as the specific individual image data, it can be additionally registered in the specific individual image data by the second correction processing by the correction means 37. The counting unit 38 counts the number of detected specific individual image data based on the image analysis processing result by the image analyzing unit 36 and the correction processing result by the correcting unit 37, and the counting result is output by the output unit 34. In the present embodiment, the first correction process (error correction of specific individual image data) and the second correction process (correction correction of specific individual image data) can be performed by the correction unit 37. It is possible to improve.

<Data structure of moving image data>
FIG. 2 is a diagram illustrating a data configuration of moving image data captured by the camera 22.
The data structure of the moving image data according to the present embodiment includes a survey ID 222, a shooting moving image 223, position information 224, date / time information 225, and shooting conditions 226. The survey ID 222 is an ID for identifying data. The captured moving image 223 is moving image data captured by the camera 22. The position information 224 is position information such as a position where a moving image is captured using GPS or the like, a flight route, and the like. The date / time information 225 is information on the date / time of shooting. The shooting conditions 226 include meteorological data such as temperature and humidity at the time of shooting, the angle of view and magnification at the time of shooting, the flight speed of the flying object 2, and the like.

<Summary of flight route>
FIG. 3 is a schematic diagram of the flight route of the flying object 2.
The flight route 51 according to this embodiment takes a point (via point) 52 on the map 5 of the area to be counted and takes the flight 2 as a starting point of flight, and the aircraft 2 takes off from the point 52 toward the point (via point) 53. When the vehicle arrives at the point 53, the route is changed to the point (route point) 54. Then, the route in which the flying object 2 flies in the direction of the arrow sequentially from the point 54, and when the vehicle 2 arrives at the point (via point) 55 that is the shooting end point, the flying object 2 returns to the point 52 that is the flight end point. And land. A base provided with a wireless charger is installed at the point 52, and the flying object 2 lands on the wireless charger. When it is detected that the flying object 2 has landed, charging of the battery of the flying object 2 is started. In this embodiment, the point 52 is a flight start and shooting start point, and is a flight end point, and the point 55 is a shooting end point. However, the present invention is not limited thereto, and the point 52 is a flight start and shooting start point. The point 55 may be the flight end point and the shooting end point. Further, the point 52 is an imaging start point, and the point 55 is an imaging end point, and the flight start point and the flight end point of the flying object 2 may be set at other positions.

<Outline of shooting method>
FIG. 4 is a schematic diagram of a method for photographing the flying object 2.
The flying object 2 flies at a higher altitude than the mountain 6 or the tree 8. As shown in FIG. 4, the camera 22 provided on the flying object 2 does not shoot directly under the flying object 2. The ground is photographed at a predetermined angle (for example, 45 degrees). This is because the shooting range per frame is wider than when shooting the ground from directly above, and when the ground is shot from directly above, the subject animal (in this case, deer 7) is hidden in the tree 8. This is because there may be cases where shooting is not possible. In addition, in order to photograph many features of the animal to be investigated, it is better to photograph at an angle than from above. In the present embodiment, the flying altitude and flying speed of the flying object 2 are constant.

<Angle of view of camera 22 of flying object 2>
FIG. 5 is an explanatory diagram showing in detail the angle of view of the camera 22 of the flying object 2. As shown in the figure, the direction of the lens 221 of the camera 22 of the flying object 2 is inclined by an angle θ with the direction from the flying object 2 toward the ground as 0 degrees. For example, the angle θ may be fixed at 30 degrees, 45 degrees, or the like, and can be changed as appropriate according to the topography of the target area, the flight altitude or the route of the aircraft 2.

<Flying control of flying object 2>
FIG. 6 is a flowchart showing a flight control flow of the flying object 2.
The flight control means 21 loads the current animal habitat research flight route data transmitted from the server 3, and controls the flight of the flying object 2 according to the loaded animal habitat research flight route.

First, in step S601, the flight control means 21 acquires current position information (latitude, longitude, etc.) of the flying object 2 from a GPS (not shown). In the present embodiment, the position information acquired in step S601 determines not only the position such as GPS but also the flight altitude.

In step S602, the flight control means 21 compares the acquired position information with the loaded flight route for animal habitat investigation, and determines whether or not it is the imaging start point set in the flight route for animal habitat investigation. If it is not the shooting start point, it will fly to the shooting start point according to the flight route set in the flight route for animal habitat investigation. In the present embodiment, since the flight start point and the imaging start point are the same point 52 as described above, the flying object 2 here rises to the altitude set by the animal habitat survey flight route.

If it is determined that the current point is the shooting start point (YES), the process proceeds to step S603, where the flight control unit 21 activates the camera 22 by the shooting control unit 23 and starts shooting a moving image.

Next, in step S604, the flight control means 21 controls the flight according to the flight route, flight altitude, and flight speed set as the animal habitat research flight route.

Next, in step S605, the flight control means 21 determines whether or not it has arrived at the photographing end point set in the flight route for animal habitat investigation. If it has arrived at the photographing end point (point 55) (YES), the process proceeds to step S606. If it has not arrived at the photographing end point (NO), the process returns to step S604 and the photographing is continued.

In step S606, the flight control means 21 ends the activation of the camera 22 by the shooting control means 23, and ends the shooting of the moving image.

In step S607, the flight control means 21 determines whether or not it has arrived at the flight end point set in the flight route for animal habitat investigation. If the flight end point has been reached (YES), the process proceeds to step S608, and the process is terminated after flying to the standby position (point 52) of the flying object 2 and landing at the standby position. In the present embodiment, after the photographing is finished, the process returns to the standby position of the flying object 2. However, the present invention is not limited to this, and the landing may be made at the end point of the flight route.

<Shooting Control of Camera 22>
FIG. 7 is a flowchart showing a shooting control flow of the camera 22.
First, in step S701, the imaging control unit 23 determines whether or not the imaging start condition is satisfied. That is, the imaging control unit 23 determines whether or not the flight control unit 21 determines that it has arrived at the flight start position in step S602.

Next, in step S701, it is determined whether or not preparation for shooting has been completed. Specifically, it is determined whether or not the angle of view, the magnification, and the like satisfy a preset condition when the camera 22 is activated in step S603. When the preparation for shooting is completed (YES), the process proceeds to step S704. On the other hand, if the preparation for photographing has not been completed (NO), the process proceeds to step S703, the camera 22 is adjusted to adjust the angle of view, the magnification, etc. so as to satisfy preset conditions, and then the process proceeds to step S704.

In step S 704, the shooting control unit 23 starts shooting a moving image with the camera 22, and sequentially stores the shot color moving image data in the storage unit 24 and the server 3 sequentially stores it in the storage unit 24. Color moving image data is sequentially transmitted.

Next, in step S705, the imaging control means 23 determines whether or not to end imaging. That is, if it is determined in step S605 that the flying object 2 has arrived at the shooting end point on the flight route, it is determined that shooting is to be ended.

Next, in step S706, the shooting control unit 23 transmits the shot moving image (moving image data) to the server 3 and ends the process. The server 3 stores the transmitted color moving image data in the DB 31 in association with the survey ID 222 as described above with reference to FIG.

<Animal habit investigation result processing>
FIG. 8 is a flowchart showing an animal habitat survey result processing flow performed by the server 3.
First, in step S801, the extraction unit 35 converts the color moving image data stored in the DB 31 into monochrome image data, and sequentially extracts still image data from the monochrome image data at a preset timing (predetermined time or predetermined frame). Extract.

Next, in step S802, the image analysis means 36 performs pattern matching with the extracted still image data with reference to the individual image information of the current counting target stored in the DB 31, and the counting target is included in the still image data. The image data having a high matching degree (similarity) with the individual image information is identified as the specific individual image data, and is sequentially stored in the DB 31 in association with the specific individual image data ID, coordinate information, and the like as the specific individual image data.

Next, in step S803, the correction unit 37 determines whether or not to execute the first correction process (error correction of specific individual image data). If the first correction process is not executed (NO), the process proceeds to step S806. If the first correction process is executed (YES), the process proceeds to step S804. In step S803, when the first correction process is not executed for a predetermined period (for example, 30 days), a first correction process promotion display that prompts the user to execute the first correction process is displayed on the monitor. The operator who has confirmed the first correction process promotion screen selects execution / non-execution of the first correction process. If execution of the first correction process is selected, the process proceeds to step S804. If non-execution of the first correction process is selected, the process proceeds to step S806.

Next, in step S804, the correction unit 37 executes a first correction process (error correction of specific individual image data). When the first correction process is executed, the next display timing of the first correction process promotion display is updated and registered at a timing after a predetermined period (for example, 30 days later) from the next day.

Next, in step S805, the registration unit 32 performs a process of deleting individual image information showing a high degree of matching with the image data removed from the specific individual image data by the first correction process from the individual image information to be counted. Do.

Next, in step S806, the correction unit 37 determines whether or not to execute the second correction process (missing correction of specific individual image data). If the second correction process is not executed (NO), the process proceeds to step S809. If the second correction process is executed (YES), the process proceeds to step S807. In step S806, if the second correction process has not been executed for a predetermined period (for example, 30 days), a second correction process promotion display that prompts the user to execute the second correction process is displayed on the monitor. The operator who confirms the second correction process promotion screen selects execution / non-execution of the second correction process. If execution of the second correction process is selected, the process proceeds to step S807. If non-execution of the first correction process is selected, the process proceeds to step S809.

Next, in step S807, the correcting unit 37 executes a second correction process (correction correction of specific individual image data). When the second correction process is executed, the next display timing of the second correction process promotion display is updated and registered at a timing after a predetermined period (for example, 30 days later) from the next day.

Next, in step S808, the registration unit 32 performs a process of additionally registering the image data newly included in the specific individual image data by the second correction process as the individual image information to be counted.

Next, in step S806, the counting unit 38 specifies the number of specific individual image data detected corresponding to the type of each animal based on the image analysis processing result by the image analysis unit 36 and the correction processing result by the correction unit 37. Count for each individual image data.

In step S807, the output unit 34 displays the count result of the counting unit 38 together with the animal habitat survey date and the name of each survey target animal and the number of animals counted on a monitor or the like, and then ends this process.

<Modified example of this embodiment>
As for the flight route, it is desirable to set the flight route, the flight altitude, the flight speed, etc. by performing the test flight of the flying object 2 during the day. Since the individual to be counted is an animal (in this embodiment, a deer or the like), the number of individuals can be counted more accurately by performing flight at night. Furthermore, by setting a flight route during the defoliation season, it is easy to count individuals without being disturbed by trees.

In the present embodiment, deer and the like are used as individuals to be counted. However, the present invention is not limited to this. For example, individual image information is stored in the DB 31 for each animal such as a frog, a horse, a bear, and a rabbit, and the still image described above is stored. Analysis processing from the data by the image analysis means 36, correction processing by the correction means 37, and registration processing by the registration means 32 may be performed. Furthermore, the individual image information is stored in the DB 31 not only for each animal but also for each category such as male, female, and child, and the above-described analysis processing by the image analysis means 36 from the still image data, correction processing by the correction means 37, and registration means. 32 may be performed.

In the present embodiment, monochrome moving image data is extracted after converting color captured moving image data to monochrome moving image data. However, the present invention is not limited to this. Color still images are extracted from color captured moving image data. After the data is extracted, it may be converted into black and white still image data. Further, the color still image data may be used to perform analysis processing by the image analysis means 36, correction processing by the correction means 37, and registration processing by the registration means 32.

Further, the camera 22 may be an infrared camera or a high sensitivity camera. In addition, when shooting is performed using an infrared camera or a high-sensitivity camera, the image may not be converted into monochrome still image data.

In the present embodiment, when the first correction process is performed by the correction unit 37, the registration unit 32 exhibits a high matching degree (for example, 85% or more) with respect to the image data excluded from the specific individual image data. The process of deleting the individual image information from the individual image information to be counted was performed. However, the present invention is not limited to this, and the individual image information is left without being deleted, and misidentification information indicating that there is an error is added. Then, update registration may be performed.

In the present embodiment, when there are a plurality of pieces of image data having a high matching degree with the individual image information to be counted in the still image data, the image data is specified as the individual image data to be counted. However, the present invention is not limited to this, and when there is at least one piece of image data having a high degree of matching with the individual image information to be counted in the still image data, this image data may be specified as the individual image data to be counted. Good.

Also, the following method can be considered as another additional registration method when the second correction process is performed. A first method is a method of additionally registering by displaying image data to be additionally registered with a mouse or the like (enclosing or clicking the image with a mouse or the like) while a still image is displayed on the monitor. As a second method, when the monitor is a touch panel display, additional registration is performed by displaying the image data to be additionally registered on the touch panel in a state where a still image is displayed (enclose or touch the image with the touch panel). how to.

In the present embodiment, the correction unit 37 executes the first correction process and the second correction process. However, the present invention is not limited to this, and either the first correction process or the second correction process is performed. May be executed only.
<Additional notes>

Preferably, in the registration step, the registered detection target animal information can be corrected based on the non-detection target image, thereby improving the accuracy of the detection target animal information. It is.

Preferably, in the correcting step, the still image is displayed, an image to be registered as the detection target animal information is received from the displayed still image, and the received image is used as the detection target animal information as the storage means. It is possible to improve the accuracy of the detection target animal information.

DESCRIPTION OF SYMBOLS 1 Animal habitat survey system 2 Aircraft 3 Server 4 Wide area communication network 21 Flight control means 22 Camera 23 Imaging | photography control means 24 Storage means 31 DB
32 Registration means 33 Input means 34 Output means 35 Extraction means 36 Image analysis means 37 Correction means 38 Counting means

Claims

A shooting step of flying a flying object including an imaging unit along a preset flight route, and shooting an imaging target area with the imaging unit;
An extraction step of extracting a still image at a predetermined timing from a captured image captured by the imaging means;
A registration step of registering detection target animal information capable of specifying an animal image to be detected in the storage means;
A detection step of detecting a detection target animal image from the still image based on the registered detection target animal information;
When the image that should not be detected as the detection target animal image is detected, a correction step of accepting correction to exclude the image from the detection target animal image;
Based on the result of the detection step and the result of the correction step, a counting step of counting the number of animal images to be detected in the imaging target area;
An output step for outputting the result of the counting,
In the registration process,
The animal habitat investigation method characterized in that the detection target animal information can be updated based on a result of the correction step.
A shooting step of flying a flying object including an imaging unit along a preset flight route, and shooting an imaging target area with the imaging unit;
An extraction step of extracting a still image at a predetermined timing from a captured image captured by the imaging means;
A registration step of registering detection target animal information capable of specifying an animal image to be detected in the storage means;
A detection step of detecting a detection target animal image from the still image based on the registered detection target animal information;
When the image to be detected as the detection target animal image is not detected, a correction step of accepting correction to include the image in the detection target animal image;
Based on the result of the detection step and the result of the correction step, a counting step of counting the number of animal images to be detected in the imaging target area;
An output step for outputting the result of the counting,
In the registration process,
The animal habitat investigation method characterized in that the detection target animal information can be updated based on a result of the correction step.
In the photographing step, the flight route can be set with a flight route and altitude so that an angle of view and a magnification at which the imaging target area is photographed by the imaging unit are constant. The animal habitat investigation method according to claim 1.
The animal habitat investigation method according to any one of claims 1 to 3, wherein the predetermined timing can be set by a predetermined time interval or a predetermined frame interval.
The animal habitat investigation method according to any one of claims 1 to 4, wherein a screen for prompting correction is displayed when the correction step is not performed for a predetermined period.
An animal habitat provided with an imaging means for imaging an imaging target area, and outputting a result of a habitat survey of an animal to be investigated based on a flying object that flies along a preset flight route and a captured image captured by the imaging means An animal habitat investigation system comprising an investigation device,
The animal habitat investigation device
Extraction means for extracting a still image at a predetermined timing from a captured image captured by the imaging means;
Registration means for registering detection target animal information capable of specifying an animal image to be detected in the storage means;
Detection means for detecting a detection target animal image from the still image based on the registered detection target animal information;
When the image that should not be detected as the detection target animal image is detected, a correction unit that receives correction to exclude the image from the detection target animal image;
Counting means for counting the number of animal images of the detection target in the imaging target area based on the result of the detection means and the result of the correction means;
Output means for outputting the result of the counting,
In the registration means,
The animal habitat investigation system characterized in that the detection target animal information can be registered based on a result of the correction means.
Extraction means for extracting a still image at a predetermined timing from a captured image captured by an imaging means for capturing an imaging target area;
Registration means for registering detection target animal information capable of specifying an animal image to be detected in the storage means;
Detection means for detecting a detection target animal image from the still image based on the registered detection target animal information;
When the image that should not be detected as the detection target animal image is detected, a correction unit that receives correction to exclude the image from the detection target animal image;
Counting means for counting the number of animal images of the detection target in the imaging target area based on the result of the detection means and the result of the correction means;
Output means for outputting the result of the counting,
In the registration means,
The animal habitat investigation device characterized in that the detection target animal information can be registered based on a result of the correction means.