An in-picture wild animal tracking and positioning method based on video monitoring, a storage medium and an equipment

Abstract

The application discloses a picture-in-picture wild animal tracking positioning method based on video monitoring, a storage medium and equipment, and comprises the following steps: setting a monocular holder camera in a nature reserve to monitor wild animals; acquiring the center coordinates on the video monitoring picture of the monocular holder camera, combining the coordinates of the grid intersection points in the video monitoring picture to determine the image rotation invariant center coordinates of the monocular holder camera; when the wild animals enter the video monitoring picture, identifying the position coordinates of the wild animals, calculating the displacement between the position coordinates of the wild animals and the image rotation invariant center coordinates of the monocular holder camera, mapping the displacement to the original picture of the monocular holder camera, converting the field angle of the monocular holder camera into the rotation angle of the monocular holder camera, controlling the rotation of the monocular holder camera, and making the wild animals always located in the center of the video monitoring picture, so that the tracking positioning of the wild animals is realized.

Description

Technical Field

[0001] This invention relates to the field of wildlife tracking and positioning technology, specifically to a method, storage medium, and device for tracking and positioning wild animals within a video surveillance frame. Background Technology

[0002] Video surveillance technology is widely used in existing biodiversity monitoring systems, especially in ecological reserves and wildlife habitats, for real-time observation and recording of wildlife behavior and activities. Traditional video surveillance systems typically employ multiple fixed or pan-tilt-zoom (PTZ) cameras deployed throughout the monitoring area to capture wildlife activity. These cameras generally use infrared, high-definition, or low-light sensors to effectively capture images under varying lighting conditions. Monitoring personnel can then use existing video analytics techniques to observe the location and activity status of wildlife.

[0003] While existing video analytics technologies can identify wildlife targets in footage, they still suffer from significant errors in accurately locating targets, especially when the targets are moving rapidly or against complex backgrounds. Furthermore, in practical applications, the unpredictable behavior of wild animals leads to insufficient efficiency and accuracy in gimbal tracking. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method, storage medium, and device for tracking and locating wild animals within video surveillance footage. This method enables rapid location of wild animals within video surveillance footage, ensuring that moving wild animals remain centered within the video surveillance frame, thus achieving rapid tracking of wild animals.

[0005] To achieve the above technical objectives, the present invention adopts the following technical solution: a method for tracking and locating wild animals within a video surveillance frame, specifically including the following steps:

[0006] Step S1: Set up a monocular pan-tilt camera in the nature reserve to monitor wild animals;

[0007] Step S2: Obtain the center coordinates on the video monitoring screen of the monocular PTZ camera, and combine them with the coordinates of the grid intersection points in the video monitoring screen to determine the rotation-invariant center coordinates of the monocular PTZ camera image.

[0008] Step S3: When a wild animal enters the video surveillance screen, identify the location coordinates of the wild animal, calculate the displacement between the location coordinates of the wild animal and the rotation-invariant center coordinates of the image of the monocular pan-tilt camera, and map the displacement onto the original image of the monocular pan-tilt camera.

[0009] Step S4: The displacement in the original image mapped to the monocular pan-tilt camera is converted into the rotation angle of the monocular pan-tilt camera through the field of view of the monocular pan-tilt camera. The monocular pan-tilt camera is rotated to keep the wild animals in the center of the video monitoring screen, so as to realize the tracking and positioning of wild animals.

[0010] Further, step S2 includes the following sub-steps:

[0011] Step S2.1: Use the center coordinates of the video monitoring screen of the monocular PTZ camera as the initial image rotation-invariant center coordinates;

[0012] Step S2.2: Construct an offset error function based on the grid intersections and the rotation-invariant center of the image in the video surveillance image;

[0013] Step S2.3: Calculate the gradient of the offset error function with respect to the rotation-invariant center of the image, and update the coordinates of the rotation-invariant center of the image;

[0014] Step S2.4: Repeat steps S2.2-S2.3 until the offset error function reaches its minimum value, and obtain the rotation-invariant center coordinates of the image from the monocular PTZ camera.

[0015] Furthermore, the process of constructing the offset error function is as follows:

[0016]

[0017] Among them, (x c ,y c R(θ) represents the rotation-invariant center coordinates of the image from the monocular pan-tilt camera, n represents the number of grid intersections in the video surveillance image, i represents the index of n, and R(θ) represents the rotation matrix of the monocular pan-tilt camera. p i This represents the offset of the i-th grid intersection point relative to the rotation-invariant center coordinates of the image. (x i ,y i ) represents the coordinates of the i-th grid intersection point, and |||| represents the norm.

[0018] Furthermore, the update process for the rotation-invariant center coordinates of the image is as follows:

[0019]

[0020] in, This represents the rotation-invariant center coordinates of the updated image, and α represents the learning rate. This represents the gradient of the offset error function with respect to the rotation-invariant center of the image.

[0021] Furthermore, the process of identifying the location coordinates of the wild animal is divided into mouse click identification and frame identification. Step S3 includes the following sub-steps:

[0022] Step S3.1: Calculate the displacement of the wild animal's position coordinates relative to the rotation-invariant center coordinates of the monocular pan-tilt camera image:

[0023] Δ x =x u -x c Δ y =y u -y c

[0024] Where, Δ x Δ represents the displacement in the horizontal direction. y Indicates displacement in the vertical direction;

[0025] Step S3.2: Based on the resolution W of the video surveillance image p ×H p Scaling the displacement:

[0026]

[0027] Where, r x r represents the proportion of horizontal displacement relative to the displacement of the video surveillance image. y It represents the proportion of vertical displacement relative to the displacement of the video surveillance image.

[0028] Step S3.3: Based on the original image resolution W of the monocular pan-tilt camera c ×H c The proportional displacement is mapped onto the raw image from the monocular gimbal camera:

[0029]

[0030] in, Indicates that r x Mapped onto the original image from the monocular pan-tilt camera, Show r y The image is mapped onto the original view from the monocular pan-tilt camera.

[0031] Furthermore, when using a frame to identify the location coordinates of wild animals, the coordinates of the frame's center are used as the location coordinates of the identified wild animals: Where (x1, y1) represents the coordinates of the top left corner of the frame, and (x2, y2) represents the coordinates of the bottom right corner of the frame.

[0032] Furthermore, the calculation process for the rotation angle of the monocular gimbal camera in step S4 is as follows:

[0033]

[0034] Where, θ h FoV represents the horizontal rotation angle of a monocular pan-tilt-zoom (PTZ) camera. h θ represents the field of view of a monocular pan-tilt camera in the horizontal direction. v This shows the vertical rotation angle of a monocular pan-tilt camera, FoV. v This indicates the field of view of a monocular pan-tilt camera in the vertical direction.

[0035] Furthermore, when using frame tracking to locate the position coordinates of wild animals, it is necessary to adjust the zoom factor of the monocular gimbal camera. The specific process is as follows:

[0036] i. Determine the size of the frame area based on the coordinates of the top left corner (x1, y1) and the bottom right corner (x2, y2):

[0037] W box =x2-x1,H box =y2-y1

[0038] Among them, W box H represents the width of the picture frame. box Indicates the height of the picture frame;

[0039] ii. Calculate the ratio of the frame area to the video surveillance screen based on the frame area size.

[0040] iii. Using the focal length-magnification relationship to determine r box Mapped to the scaling factor m = f -1 (r box ), which controls the adjustment of the zoom factor of the monocular PTZ camera, where f() represents a function of the focal length-magnification relationship.

[0041] Furthermore, the present invention also provides a computer-readable storage medium storing a computer program that enables a computer to execute the aforementioned method for tracking and locating wild animals within video surveillance footage.

[0042] Furthermore, the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the aforementioned method for tracking and locating wild animals within a video surveillance frame.

[0043] Compared with existing technologies, the present invention has the following beneficial effects: The wildlife tracking and positioning method based on video surveillance in this invention constructs an offset error function between the grid intersection points in the video surveillance image and the rotation-invariant center of the monocular PTZ camera image. Combined with the gradient descent method, it optimizes the rotation-invariant center of the monocular PTZ camera image, thereby quickly finding the rotation-invariant center of the monocular PTZ camera image, providing a solid foundation for subsequent accurate tracking of wildlife in video footage. At the same time, the present invention calculates the displacement between the rotation-invariant center of the monocular PTZ camera image and the position of the wildlife in the video surveillance image in real time, maps the displacement to the original image of the monocular PTZ camera, guides the angle adjustment of the monocular PTZ camera, thereby locking the position of the wildlife and ensuring that the wildlife is always in the center of the video surveillance image, improving the efficiency and accuracy of wildlife tracking. Attached Figure Description

[0044] Figure 1 This is a flowchart of the method for tracking and locating wild animals within a video surveillance frame according to the present invention. Detailed Implementation

[0045] The technical solution of the present invention will be further explained and described below with reference to the accompanying drawings.

[0046] like Figure 1 This is a flowchart of the wildlife tracking and positioning method based on video surveillance footage according to the present invention. The wildlife tracking and positioning method specifically includes the following steps:

[0047] Step S1: Set up a monocular pan-tilt camera in the nature reserve to monitor wild animals.

[0048] Step S2: The video monitoring screen of the monocular pan-tilt camera is displayed on the player. Due to the optical distortion, installation angle error, and cropping of the monocular pan-tilt camera, the rotation-invariant center of the image captured by the monocular pan-tilt camera may deviate from the diagonal center of the displayed video monitoring screen. Therefore, it is necessary to optimize the center of the video monitoring screen and find the rotation-invariant center of the monocular pan-tilt camera image to provide a solid foundation for accurate tracking of wild animals. This invention determines the coordinates of the rotation-invariant center of the monocular pan-tilt camera image by obtaining the center coordinates of the video monitoring screen and combining them with the coordinates of the grid intersection points in the video monitoring screen; including the following sub-steps:

[0049] Step S2.1: Use the center coordinates of the video monitoring screen of the monocular PTZ camera as the initial image rotation-invariant center coordinates;

[0050] Step S2.2: Construct an offset error function based on the grid intersections and the rotation-invariant center of the image in the video surveillance image to quantify the deviation between the rotation-invariant center of the image and the known position;

[0051] The process of constructing the offset error function in this invention is as follows:

[0052]

[0053] Among them, (x c ,y c R(θ) represents the rotation-invariant center coordinates of the image from the monocular pan-tilt camera, n represents the number of grid intersections in the video surveillance image, i represents the index of n, and R(θ) represents the rotation matrix of the monocular pan-tilt camera. p i This represents the offset of the i-th grid intersection point relative to the rotation-invariant center coordinates of the image. (x i ,y i ) represents the coordinates of the i-th grid intersection point, and |||| represents the norm.

[0054] Step S2.3: Calculate the gradient of the offset error function with respect to the rotation-invariant center of the image, and update the coordinates of the rotation-invariant center of the image.

[0055]

[0056] in, This represents the rotation-invariant center coordinates of the updated image, and α represents the learning rate. This represents the gradient of the offset error function with respect to the rotation-invariant center of the image.

[0057] Step S2.4: Repeat steps S2.2-S2.3 until the offset error function reaches its minimum value, and obtain the rotation-invariant center coordinates of the image from the monocular PTZ camera, thus enabling rapid convergence to the optimal solution.

[0058] Step S3: When a wild animal enters the video surveillance frame, its position coordinates are identified. The displacement between the wild animal's position coordinates and the rotation-invariant center coordinates of the monocular pan-tilt camera's image is calculated. This displacement is mapped onto the original image of the monocular pan-tilt camera. Since the rotation-invariant center of the monocular pan-tilt camera's image is a stable and unchanging reference point, it is unaffected by the rotation or tilt of the monocular pan-tilt camera, enabling precise positioning of the wild animal within the monocular pan-tilt camera's field of view. The process of identifying the wild animal's position coordinates is divided into mouse click recognition and frame recognition, specifically including the following sub-steps:

[0059] Step S3.1: Calculate the displacement of the wild animal's position coordinates relative to the rotation-invariant center coordinates of the monocular pan-tilt camera image:

[0060] Δ x =x u -x c Δ y =y u -y c

[0061] Where, Δ x Δ represents the displacement in the horizontal direction. y Indicates displacement in the vertical direction;

[0062] When using a frame to identify the location coordinates of wild animals, the coordinates of the frame's center are used as the location coordinates of the identified wild animals. Where (x1, y1) represents the coordinates of the top left corner of the frame, and (x2, y2) represents the coordinates of the bottom right corner of the frame.

[0063] Step S3.2: Based on the resolution W of the video surveillance image p ×H p Scaling the displacement:

[0064]

[0065] Where, r x r represents the proportion of horizontal displacement relative to the displacement of the video surveillance image. y It represents the proportion of vertical displacement relative to the displacement of the video surveillance image.

[0066] Step S3.3: Based on the original image resolution W of the monocular pan-tilt camera c ×H c The proportional displacement is mapped onto the raw image from the monocular gimbal camera:

[0067]

[0068] in, Indicates that r x Mapped onto the original image from the monocular pan-tilt camera, Show r y The image is mapped onto the original view from the monocular pan-tilt camera.

[0069] Step S4: The displacement in the original image mapped to the monocular pan-tilt camera is converted into the rotation angle of the monocular pan-tilt camera through the field of view of the monocular pan-tilt camera. The rotation of the monocular pan-tilt camera is controlled so that the wild animal is always in the center of the video monitoring screen, thereby realizing the tracking and positioning of wild animals and improving the efficiency and accuracy of wild animal tracking.

[0070] The calculation process for the rotation angle of the monocular gimbal camera in this invention is as follows:

[0071]

[0072] Where, θ h FoV represents the horizontal rotation angle of a monocular pan-tilt-zoom (PTZ) camera. h θ represents the field of view of a monocular pan-tilt camera in the horizontal direction. v This shows the vertical rotation angle of a monocular pan-tilt camera, FoV. v This indicates the field of view of a monocular pan-tilt camera in the vertical direction.

[0073] When using frame tracking to locate the coordinates of wild animals, the zoom level of the monocular pan-tilt camera needs to be adjusted based on the actual size and distance of the animal to ensure it remains clear and easily identifiable in the video surveillance footage. Increasing the zoom level magnifies the image of the wild animal, improving clarity; while decreasing the zoom level expands the field of view, facilitating the search and tracking of wild animals in a wider scene.

[0074] The specific process for adjusting the zoom factor of a monocular gimbal camera in this invention is as follows:

[0075] i. Determine the size of the frame area based on the coordinates of the top left corner (x1, y1) and the bottom right corner (x2, y2):

[0076] W box =x2-x1,H box =y2-y1

[0077] Among them, W box H represents the width of the picture frame. box Indicates the height of the picture frame;

[0078] ii. Calculate the ratio of the frame area to the video surveillance screen based on the frame area size.

[0079] iii. Using the focal length-magnification relationship to determine r box Mapped to the scaling factor m = f -1 (r box ), which controls the adjustment of the zoom factor of the monocular PTZ camera, where f() represents a function of the focal length-magnification relationship.

[0080] In one technical solution of the present invention, the focal length-magnification relationship of the monocular gimbal camera is determined by a constructed dynamic differential equation:

[0081]

[0082] Where f(t) represents the focal length as a function of time t, M(t) represents the optical magnification as a function of time t, τ(t) represents the motor drive input parameters as a function of time t, ∈(t) represents the noise as a function of time t, and g() represents the nonlinear function.

[0083] In one technical solution of the present invention, a computer-readable storage medium is also provided, storing a computer program that enables a computer to execute the aforementioned method for tracking and locating wild animals within video surveillance footage.

[0084] In one technical solution of the present invention, an electronic device is also provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned method for tracking and locating wild animals within a video surveillance frame.

[0085] The method for tracking and locating wild animals within video surveillance footage, as described in this invention, can also be applied to tracking and locating any moving object within a video surveillance frame.

[0086] In the embodiments disclosed in this application, a computer storage medium may be a tangible medium that may contain or store programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of computer storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0087] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0088] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for tracking and locating wild animals within video surveillance footage, characterized in that, Specifically, the steps include the following: Step S1: Set up a monocular pan-tilt camera in the nature reserve to monitor wild animals; Step S2: Obtain the center coordinates of the video surveillance screen from the monocular PTZ camera, and combine them with the coordinates of the grid intersection points in the video surveillance screen to determine the rotation-invariant center coordinates of the monocular PTZ camera image; this includes the following sub-steps: Step S2.1: Use the center coordinates of the video monitoring screen of the monocular PTZ camera as the initial image rotation-invariant center coordinates; Step S2.2: Construct an offset error function based on the grid intersections and the rotation-invariant center of the image in the video surveillance footage: in, This represents the rotation-invariant center coordinates of the image from a monocular pan-tilt-zoom (PTZ) camera. n This indicates the number of grid intersections in the video surveillance footage. i express n index, This represents the rotation matrix of a monocular pan-tilt-zoom (PTZ) camera. ; Indicates the first i The offset of each grid intersection point relative to the rotation-invariant center coordinates of the image. , Indicates the first i The coordinates of the grid intersections, where || represents the norm; Step S2.3: Calculate the gradient of the offset error function with respect to the rotation-invariant center of the image, and update the coordinates of the rotation-invariant center of the image; Step S2.4: Repeat steps S2.2-S2.3 until the offset error function reaches its minimum value, and obtain the rotation-invariant center coordinates of the image from the monocular PTZ camera. Step S3: When a wild animal enters the video surveillance screen, identify the location coordinates of the wild animal, calculate the displacement between the location coordinates of the wild animal and the rotation-invariant center coordinates of the image of the monocular pan-tilt camera, and map the displacement onto the original image of the monocular pan-tilt camera. Step S4: The displacement in the original image mapped to the monocular pan-tilt camera is converted into the rotation angle of the monocular pan-tilt camera through the field of view of the monocular pan-tilt camera. The monocular pan-tilt camera is rotated to keep the wild animals in the center of the video monitoring screen, so as to realize the tracking and positioning of wild animals.

2. The method for tracking and locating wild animals within a video surveillance frame according to claim 1, characterized in that, The process of updating the rotation-invariant center coordinates of the image is as follows: in, This represents the rotation-invariant center coordinates of the updated image. Indicates the learning rate. The ) represents the gradient of the offset error function with respect to the rotation-invariant center of the image. , .

3. The method for tracking and locating wild animals within a video surveillance frame according to claim 2, characterized in that, The process of identifying the location coordinates of the wild animals is divided into mouse click identification and frame identification. Step S3 includes the following sub-steps: Step S3.1: Calculate the displacement of the wild animal's position coordinates relative to the rotation-invariant center coordinates of the monocular pan-tilt camera image: ， in, Indicates displacement in the horizontal direction. Indicates displacement in the vertical direction; Step S3.2: Based on the resolution of the video surveillance image Scaling the displacement: ， in, This indicates the proportion of horizontal displacement relative to the displacement of the video surveillance image. It represents the proportion of vertical displacement relative to the displacement of the video surveillance image. Step S3.3: Based on the original image resolution of the monocular pan-tilt camera The proportional displacement is mapped onto the raw image from the monocular gimbal camera: ， in, Indicates will Mapped onto the original image from the monocular pan-tilt camera, Showing The image is mapped onto the original view from the monocular pan-tilt camera.

4. The method for tracking and locating wild animals within a video surveillance frame according to claim 3, characterized in that, When using a frame to identify the location coordinates of wild animals, the coordinates of the frame's center are used as the location coordinates of the identified wild animals: , ,in,( , () represents the coordinates of the top left corner of the frame. , ) represents the coordinates of the bottom right corner of the frame.

5. A method for tracking and locating wild animals within a video surveillance frame according to claim 3, characterized in that, The calculation process for the rotation angle of the monocular pan-tilt camera in step S4 is as follows: ， in, This indicates the rotation angle of the monocular pan-tilt camera in the horizontal direction. This indicates the field of view of a monocular pan-tilt-zoom (PTZ) camera in the horizontal direction. This shows the vertical rotation angle of a monocular pan-tilt-zoom camera. This indicates the field of view of a monocular pan-tilt camera in the vertical direction.

6. A method for tracking and locating wild animals within a video surveillance frame according to claim 5, characterized in that, When using frame tracking to locate the position coordinates of wild animals, it is necessary to adjust the zoom factor of the monocular gimbal camera. The specific process is as follows: i. Based on the coordinates of the top left corner of the drawing frame ( , ) and the coordinates of the lower right corner ( , Determine the dimensions of the picture frame area: ， in, Indicates the width of the picture frame. Indicates the height of the picture frame; ii. Calculate the ratio of the frame area to the video surveillance screen based on the frame area size. ; iii. Using the focal length-magnification relationship to... Mapped to scaling factor Controls the adjustment of the zoom factor of the monocular pan-tilt camera, among which, f ( ) represents a function representing the focal length-magnification relationship.

7. A computer-readable storage medium storing a computer program, characterized in that, The computer program causes the computer to execute the video surveillance-based wildlife tracking and location method as described in any one of claims 1-6.

8. An electronic device, characterized in that, include: The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the video surveillance-based wildlife tracking and positioning method as described in any one of claims 1-6.

Images