A bionic field of view construction method based on physiological bone structure characteristics of animals

Abstract

The application discloses a kind of based on the bionic field of view construction method of animal physiological skeleton structure characteristics, belong to the technical field of the intersection of computer vision and animal behavior, including: establishing coordinate system system;Obtain the physiological skeleton structure parameter of animal head and neck, establish the kinematic model from neck base coordinate system through each neck cervical joint coordinate system to left and right eye socket coordinate system;Obtain the physiological structure of eyeball, construct binocular visual field parameterization model on the basis of kinematic model;The pose transformation matrix of smart necklace worn on the neck of animal is obtained, the rotation angle of each cervical joint is solved by inverse kinematics method, and the pose transformation matrix of left and right eye socket coordinate system in neck base coordinate system is obtained by substituting into kinematic equation and converted to geodetic coordinate system;The pose transformation matrix of left and right eye socket coordinate system in geodetic coordinate system is fused with binocular visual field parameterization model, and the bionic field of view of animal in three-dimensional space is reconstructed. It can accurately reconstruct the first visual angle visual field of animal in natural state.

Description

Technical Field

[0001] This invention belongs to the technical field of computer vision and animal behavior, specifically relating to a biomimetic field of view construction method based on the physiological skeletal structure characteristics of animals. Background Technology

[0002] In studying animal herding behavior, the impact of visual perception on behavior is often analyzed. Domestic scholars have primarily studied the effects of light on the vision and physiology of poultry, aiming to provide welfare guarantees and effective interventions for poultry farming and production. Ballerini et al. studied how starlings maintain a minimum distance in flocks that is commensurate with their wingspan, demonstrating that visual perception plays a crucial role in avoiding collisions and maintaining appropriate distances during flight. Strandburg-Peshkin's research showed the directional selection and compromise mechanisms of baboon groups towards multiple movement initiators, revealing a direct relationship between visual perception and herding behavior. These studies demonstrate the direct impact of an animal's field of vision and visual field on the animal itself and its herding behavior.

[0003] Currently, there are two main categories of methods for obtaining animal field of view and field of vision: (1) Calculating the animal's field of view and field of vision by using the field of vision captured by cameras deployed in the environment and the orientation of the animal's eyes. This method is costly and complex. (2) Simulating the position of the animal's head and eyes, installing cameras, and capturing the field of vision. However, for real animal herd movement research, it is not practical to install cameras in the animal's eyes, as it would interfere with the animal's behavior. This method is not feasible and can only be used in the laboratory as a verification auxiliary measure. Summary of the Invention

[0004] The purpose of this invention is to address the above-mentioned shortcomings in the prior art by providing a biomimetic field of view construction method based on the physiological skeletal structure characteristics of animals, thereby solving the technical problem that it is difficult to accurately obtain the first-person field of view of animals in a natural state in the prior art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A biomimetic field-of-view construction method based on animal physiological skeletal structure features includes: Establish a coordinate system, which includes a geodetic coordinate system, a neck-based coordinate system, a neck-cervical vertebral joint coordinate system, and a left and right orbital coordinate system; Obtain physiological skeletal structure parameters of the animal's head and neck, and establish a kinematic model based on the parameters. The kinematic model is described by kinematic equations from the neck base coordinate system to the left and right orbit coordinate systems, and the kinematic equations pass through the coordinate systems of each neck vertebra joint in sequence. Obtain the physiological structure of the eyeball, and construct a binocular visual field parameterization model based on the kinematic model; Obtain the pose transformation matrix of the smart collar worn around the animal's neck, and solve the rotation angle of each cervical vertebra joint using inverse kinematics. Substitute the solved rotation angles of each cervical vertebra joint into the kinematic equation to obtain the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system. Transform the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system to the geodetic coordinate system to obtain the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system. The pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system is fused with the binocular visual field parameterization model to reconstruct the biomimetic visual field of the animal in three-dimensional space.

[0006] In some embodiments, the head and neck skeletal parameters include cervical vertebral segment dimensions and articular surface angle range of motion, and the kinematic equations are constructed in the following manner: Obtain the cervical vertebral segment dimensions and articular surface angle range of motion; Analyze the characteristics of neck-head motion degrees of freedom to determine the vertical and horizontal rotation angles of each cervical vertebrae. A kinematic equation is constructed from the neck coordinate system to the left and right orbital coordinate systems. This kinematic equation describes the motion relationships between the cervical vertebrae through a homogeneous transformation matrix. The kinematic equation is expressed as follows: ; In the formula, This represents the homogeneous pose transformation matrix from the neck base coordinate system to the left and right eye socket coordinate systems; Indicates multiplication; From the first The cervical vertebrae to the first The homogeneous pose transformation matrix of each cervical vertebra represents the motion relationship between adjacent cervical vertebral segments. It is the first The vertical rotation angle of each cervical vertebra; It is the first The horizontal rotation angle of the cervical vertebrae; From the first A fixed transformation matrix from the cervical spine to the left and right eyes.

[0007] In some embodiments, the homogeneous transformation matrix between cervical vertebrae segments is: ; In the formula, It is the first The length of each cervical vertebral segment.

[0008] In some embodiments, the ocular physiological structure includes the eyeball radius, visual axis angle, and eyeball rotation angle range. The step of acquiring the ocular physiological structure, based on the kinematic model, and constructing a binocular visual field parameterized model, includes: Based on the kinematic model, the initial pose of the left and right orbital coordinate systems in the neck base coordinate system is determined. The position of the eyeball center in the orbital coordinate system is determined based on the eyeball radius, and the vertex of the field cone is set to be located at the eyeball center; The relative direction of the visual axes of the left and right eyes is determined by the angle between the visual axes; Based on the range of eyeball rotation, the boundary angles of the field cones for the left and right eyes are determined respectively, and a monocular field cone constrained by the boundary angles with the visual axis as the central axis is constructed in the orbital coordinate system. Transform the left and right monocular field-of-view cones to the same coordinate system, and obtain the binocular overlap region by finding the intersection through spatial geometry. Based on overlapping and non-overlapping regions, a parameterized model containing field-of-view boundary curves and sensitivity distribution information is generated.

[0009] In some embodiments, the field of view is divided into multiple functional zones based on the field of view boundary curve and sensitivity distribution information; wherein the functional zones include a central binocular fusion zone, a binocular fusion transition zone, a unilateral eye field of view zone, and a peripheral low-sensitivity zone.

[0010] In some embodiments, obtaining the pose transformation matrix of the smart collar worn around the animal's neck and solving the rotation angles of each cervical vertebrae using inverse kinematics methods includes: Based on the pose transformation matrix of the smart collar, the cervical spine motion parameter vector is solved using a nonlinear optimization method: ; In the formula, It is a cervical spine motion parameter vector, which includes the rotation angles of all cervical vertebrae; It is a pose transformation matrix obtained by monitoring the smart collar worn around the neck; It is the calibration offset matrix; It is the Frobenius norm.

[0011] In some embodiments, the step of transforming the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system to the geodetic coordinate system to obtain the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system includes: Obtain the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system. ; The pose transformation matrix of the left and right eye socket coordinate systems in the neck base coordinate system. Multiply the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system The pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system is obtained. : .

[0012] In some embodiments, fusing the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system with the binocular visual field parameterization model to reconstruct the animal's biomimetic visual field in three-dimensional space includes: Based on the left and right eye field conic geometric parameters and functional partition boundaries defined in the binocular field parametric model, a field geometry fixed to the left and right eye sockets is established in virtual three-dimensional space. Obtain the pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system in real time; The field-of-view geometry is transformed to the geodetic coordinate system according to the corresponding pose transformation matrix to obtain the position, orientation and coverage area of ​​the left and right fields of view in the geodetic coordinate system. The transformed left and right fields of view are spatially superimposed to generate a complete dynamic bionic field of view that includes the binocular overlap area and the monocular area.

[0013] In some embodiments, it also includes: A behavior-viewpoint mapping model is established. By analyzing the variation patterns of viewpoint parameters under different behavior states, a mapping function is constructed that takes behavior state parameters, environmental state parameters, and internal state parameters as inputs and viewpoint distribution parameters as outputs. ; In the formula, It is a moment The field of view distribution parameter vector includes the field of view coverage, center position, and sensitivity distribution region characteristics; These are behavioral state parameters, determined by the behavioral classification algorithm; These are environmental state parameters that describe the characteristics of the surrounding environment; These are internal state parameters that reflect the animal's physiological and cognitive state; It is a mapping function.

[0014] The present invention provides a biomimetic field-of-view construction method based on the physiological skeletal structure characteristics of animals, which has the following beneficial effects: By using a smart collar worn around the neck and combining kinematic equations constructed based on real physiological skeletal structure features, the pose transformation matrix of the eyes in three-dimensional space is accurately calculated. This pose transformation matrix is ​​then fused with a binocular visual field parameterization model to reconstruct the animal's biomimetic visual field in three-dimensional space. This method eliminates the need for devices installed in the animal's eyes, minimizing interference with the animal's natural behavior, while ensuring the authenticity and continuity of data acquisition. It solves the technical problem of accurately acquiring the first-person field of view of animals in natural conditions, providing a practical and feasible technical means for field ecological monitoring, animal behavior research, intelligent grazing supervision, and wildlife conservation. Attached Figure Description

[0015] Figure 1 This is a flowchart of a biomimetic field of view construction method based on the physiological skeletal structure characteristics of animals according to the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0018] If the application documents contain similar descriptions such as "first, second, third", the following explanation shall be added: In the following description, the terms "first, second, third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] To address the problems existing in related technologies, this application provides a biomimetic field-of-view construction method based on animal physiological skeletal structural features. The executing entity of this method can be an electronic device. The electronic device can be various types of terminals such as laptops, tablets, desktop computers, set-top boxes, and mobile devices (e.g., mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), or it can be implemented as a server. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0021] In some embodiments, the functions implemented by the construction method provided in this application can be achieved by the processor of an electronic device calling program code, wherein the program code can be stored in a computer storage medium.

[0022] Example 1

[0023] This embodiment provides a biomimetic field of view construction method based on the physiological skeletal structure characteristics of animals. Figure 1 This is a flowchart of a biomimetic field-of-view construction method based on animal physiological skeletal structure features according to the present invention, such as... Figure 1 As shown, it includes: Step S1: Establish a coordinate system, which includes a geodetic coordinate system, a neck base coordinate system, a neck vertebral joint coordinate system, and a left and right orbital coordinate system; In this embodiment of the invention, to accurately describe the motion transmission process from the animal's body to its eyeballs, a coordinate system is first established, all following the right-hand rule: The Earth coordinate system {Global}, constructed within the animal's environment, is a fixed coordinate system whose pose remains unchanged throughout the process, used to describe the animal's absolute pose (position and posture) within it. The neck base coordinate system {Neck0}, constructed at the junction of the animal's torso and neck (i.e., the junction of the first thoracic vertebra or the last cervical vertebra with the torso), is a moving coordinate system. Its pose changes relative to the Earth coordinate system {Global} as the animal's torso moves. The cervical vertebra joint coordinate system {Necki}, constructed at the center of the joint of the i-th cervical vertebra segment, changes its pose as the joint moves. The left and right eye socket coordinate systems {EyeL} and {EyeR}, constructed at the centers of the left and right eye sockets, are moving coordinate systems whose pose is determined by both the animal's torso movement and neck movement. The pose information of each coordinate system is described by a homogeneous pose transformation matrix. The position is the three-dimensional coordinates (X, Y, Z) of the origin of the coordinate system, and the pose is the direction cosine of the X, Y, Z axes of the coordinate system relative to the parent coordinate system.

[0024] Step S2: Obtain the physiological skeletal structure parameters of the animal's head and neck, and establish a kinematic model based on the parameters. The kinematic model is described by kinematic equations from the neck base coordinate system to the left and right orbit coordinate systems. The kinematic equations pass through the coordinate systems of each neck vertebra joint in sequence. In this embodiment of the invention, the aim is to establish the motion transmission relationship from the base of the neck to the eye sockets in an animal. Three-dimensional scanning reverse reconstruction techniques, such as CT, laser scanning, and structured light scanning, are used to obtain precise three-dimensional parameters of the animal's neck-cranial skeleton system. A kinematic model is then constructed based on these parameters. The kinematic model is described by kinematic equations from the neck base coordinate system to the left and right eye socket coordinate systems. These kinematic equations sequentially pass through the coordinate systems of each cervical vertebra joint, accurately depicting the motion transmission relationship between the segments of the cervical vertebrae.

[0025] In some embodiments, the head and neck skeletal parameters include cervical vertebral segment dimensions and articular surface angle range of motion, and the kinematic equations are constructed in the following manner: Step S21: Obtain the cervical vertebral segment dimensions and articular surface angle range of motion; Step S22: Analyze the characteristics of neck-head motion freedom and determine the vertical and horizontal rotation angles of each cervical vertebrae. Step S23: Construct the kinematic equations from the neck base coordinate system to the left and right orbit coordinate systems. The kinematic equations describe the motion relationships between the segments of the cervical spine through a homogeneous transformation matrix. The kinematic equations are expressed as follows: ; In the formula, This represents the homogeneous pose transformation matrix from the neck base coordinate system to the left and right eye socket coordinate systems; Indicates multiplication; From the first The cervical vertebrae to the first The homogeneous pose transformation matrix of each cervical vertebra represents the motion relationship between adjacent cervical vertebral segments. It is the first The vertical rotation angle of each cervical vertebra; It is the first The horizontal rotation angle of the cervical vertebrae; From the first A fixed transformation matrix from the cervical spine to the left and right eyes.

[0026] In this embodiment of the invention, the length of each cervical vertebra segment is measured using a kinematic model. Taking yaks as an example, they typically have seven cervical vertebrae, that is... Simultaneously, the physiological range of motion of each joint surface was measured, including vertical rotation angle (pitch angle) and horizontal rotation angle (yaw angle). Then, the vertical rotation angle range of each cervical vertebra joint was calculated. Calculate the range of horizontal rotation angles for each cervical vertebra joint. Based on this, each cervical vertebra joint is simplified into a model with two rotational degrees of freedom. Finally, kinematic equations are constructed from the neck base coordinate system to the left and right orbital coordinate systems, and homogeneous transformation matrices are used to describe the pose relationship between adjacent coordinate systems.

[0027] In some embodiments, the homogeneous transformation matrix between cervical vertebrae segments is: ; In the formula, It is the first The length of each cervical vertebral segment.

[0028] Understandably, given the specific vertebral segment length of an individual animal... Real-time movement angles of each neck joint , and the fixed transformation matrix of head-eye This allows for the accurate calculation of the pose matrix of the left and right eye socket coordinate systems in the neck base coordinate system.

[0029] Step S3: Obtain the physiological structure of the eyeball, and construct a binocular visual field parameterization model based on the kinematic model; In some embodiments, step S3 includes: Step S31: Based on the kinematic model, determine the initial pose of the left and right orbital coordinate systems in the neck base coordinate system; Step S32: Determine the position of the eyeball center in the orbital coordinate system based on the eyeball radius, and set the vertex of the field cone to be located at the eyeball center; Step S33: Determine the relative direction of the visual axes of the left and right eyes based on the angle between the visual axes; Step S34: Based on the range of eyeball rotation, determine the boundary angles of the left and right eye field cones respectively, and construct a monocular field cone with the visual axis as the central axis and constrained by the boundary angles in the orbital coordinate system; Step S35: Transform the left and right monocular field-of-view cones to the same coordinate system, and obtain the binocular overlap region by finding the intersection through spatial geometry; Step S36: Generate a parameterized model containing field-of-view boundary curves and sensitivity distribution information based on overlapping and non-overlapping regions.

[0030] In this embodiment of the invention, the aim is to construct a static field-of-view perception model based on physiological structure. Precise three-dimensional morphological data of the animal's eyeball is obtained through 3D scanning, and then a parametric model of the head-bone-eyeball is reconstructed based on the kinematic model. First, based on the kinematic model, the initial pose of the left and right orbital coordinate systems in the neck base coordinate system is determined, providing a spatial reference for subsequent field-of-view construction. Second, the position of the eyeball center in the orbital coordinate system is determined according to the eyeball radius, and the vertex of the field-of-view cone is set at the eyeball center, serving as the geometric starting point of the monocular field of view. Then, the relative direction of the left and right eye visual axes is determined according to the visual axis angle, reflecting the natural orientation characteristics of the animal's eyes. Next, based on the eyeball rotation angle range, the boundary angles of the left and right eye field-of-view cones are determined respectively, and a monocular field-of-view cone constrained by the boundary angles and centered on the visual axis is constructed in the orbital coordinate system, simplifying the field of view of each eye into a cone starting from the eyeball center with a certain solid angle. Subsequently, the left and right monocular field-of-view conic shapes are transformed to the same coordinate system, and the binocular overlapping region is obtained through spatial geometric intersection, thus clarifying the range of joint perception by both eyes. Finally, based on the overlapping and non-overlapping regions, a binocular field-of-view parameterized model containing field-of-view boundary curves and sensitivity distribution information is generated, providing a foundation for subsequent field-of-view functional zoning and dynamic reconstruction.

[0031] In some embodiments, the field of view is divided into multiple functional zones based on the field of view boundary curve and sensitivity distribution information; wherein the functional zones include a central binocular fusion zone, a binocular fusion transition zone, a unilateral eye field of view zone, and a peripheral low-sensitivity zone.

[0032] In this embodiment of the invention, to more realistically reflect visual perception characteristics, the entire binocular visual field is divided into multiple functional zones based on the distribution pattern of photoreceptor cells on the retina and the visual field boundary curve and sensitivity distribution information: The central binocular fusion zone, the central part of the overlapping binocular visual fields, possesses the highest visual acuity and stereoscopic vision, and is the core area for fine perception and fixation in animals; the binocular fusion transition zone, the area where binocular visual fields overlap but are not central, still possesses stereoscopic vision, but with lower visual acuity than the central binocular fusion zone; the unilateral visual field zone, the area perceived by only one eye, lacks stereoscopic vision and is mainly used for monitoring the surrounding environment and detecting motion; and the peripheral low-sensitivity zone, the outermost area of ​​the visual field, has the lowest visual sensitivity and is mainly used for coarse environmental perception and early warning. These functional zones enable the model to more realistically reflect the visual perception characteristics of animals.

[0033] Step S4: Obtain the pose transformation matrix of the smart collar worn around the animal's neck, and solve for the rotation angle of each cervical vertebra joint using inverse kinematics; substitute the solved rotation angles of each cervical vertebra joint into the kinematic equation to obtain the pose transformation matrix of the left and right eye socket coordinate systems in the neck base coordinate system; transform the pose transformation matrix of the left and right eye socket coordinate systems in the neck base coordinate system to the geodetic coordinate system to obtain the pose transformation matrix of the left and right eye socket coordinate systems in the geodetic coordinate system. In this embodiment of the invention, a customized smart collar is worn around the animal's neck. This collar can incorporate a high-precision inertial measurement unit (IMU) and a global positioning module (GPS). The IMU includes a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer for real-time measurement of the collar's attitude and motion; the GPS is used to acquire the animal's geographic location information. Through data fusion algorithms such as Kalman filtering, the IMU and GPS data are combined to calculate the pose transformation matrix of the collar's coordinate system relative to the geodetic coordinate system in real time. The data is transmitted in real time to the data processing center via wireless communication modules (such as 4G or LoRa). Initially, the animal needs to maintain a standard posture (such as standing naturally with its head horizontal and forward), the smart collar's pose is recorded, and the fixed offset matrix between the smart collar's coordinate system and the neck's base coordinate system, i.e., the calibration offset matrix, is calculated. .

[0034] In some embodiments, obtaining the pose transformation matrix of the smart collar worn around the animal's neck and solving the rotation angles of each cervical vertebrae using inverse kinematics methods includes: Based on the pose transformation matrix of the smart collar, the cervical spine motion parameter vector is solved using a nonlinear optimization method: ; In the formula, It is a cervical spine motion parameter vector, which includes the rotation angles of all cervical vertebrae; It is a pose transformation matrix obtained by monitoring the smart collar worn around the neck; It is the calibration offset matrix; It is the Frobenius norm.

[0035] In some embodiments, the step of transforming the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system to the geodetic coordinate system to obtain the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system includes: Obtain the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system. ; The pose transformation matrix of the left and right eye socket coordinate systems in the neck base coordinate system. Multiply the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system The pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system is obtained. : .

[0036] In this embodiment of the invention, firstly, based on the pose transformation matrix of the smart collar, the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system is calculated. Secondly, since the smart collar is worn externally on the neck, it is necessary to solve for the real-time angles of the joints inside the neck, which is an inverse kinematics problem. A nonlinear optimization method is used to solve this problem, constructing an objective function and finding a set of optimal cervical spine motion parameter vectors. Since the number of cervical vertebrae is limited, this optimization problem can be solved quickly and in real time using algorithms such as Levenberg-Marquardt. Then, the optimized cervical spine motion parameter vector is... Substituting the constructed kinematic equations, the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system is calculated. Then, transform it to the geodetic coordinate system. At this point, the pose transformation matrix of the left and right eye socket coordinate systems in the geodetic coordinate system is obtained. .

[0037] Step S5: Fuse the pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system with the binocular visual field parameterization model to reconstruct the animal's biomimetic visual field in three-dimensional space.

[0038] In some embodiments, step S5 includes: Step S51: Based on the left and right eye field conic geometric parameters and functional partition boundaries defined in the binocular field parametric model, establish the field geometry that is fixed to the left and right eye sockets in the virtual three-dimensional space. Step S52: Obtain the pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system, which is calculated in real time; Step S53: Transform the field of view geometry to the geodetic coordinate system according to the corresponding pose transformation matrix to obtain the position, orientation and coverage area of ​​the left and right fields of view in the geodetic coordinate system; Step S54: Spatially superimpose the transformed left and right fields of view to generate a complete dynamic bionic field of view that includes the binocular overlap area and the monocular area.

[0039] In this embodiment of the invention, the binocular field-of-view parameterization model is a static geometric model defined in the left and right orbital coordinate systems. Pose transformation matrix. This is a homogeneous transformation matrix from the left and right orbital coordinate systems to the geodetic coordinate system, defining the position of the orbits in the geodetic system and the orientation of the three coordinate axes. First, based on the left and right eye field-of-view conical geometric parameters and functional partition boundaries defined in the binocular field-of-view parameterization model, a field-of-view geometry fixed to the left and right orbits is established in virtual 3D space. Then, the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system, calculated in real-time, is obtained. Next, the field-of-view geometry is transformed to the geodetic coordinate system according to the corresponding pose transformation matrix, obtaining the position, orientation, and coverage area of ​​the left and right eye fields in the geodetic coordinate system. Finally, the transformed left and right eye fields are spatially superimposed to generate a complete dynamic bionic field of view containing the binocular overlap area and the monocular area, thereby realizing the dynamic reconstruction of the bionic field of view of the animal in 3D space.

[0040] Example 2

[0041] Based on Embodiment 1, it also includes: A behavior-viewpoint mapping model is established. By analyzing the variation patterns of viewpoint parameters under different behavior states, a mapping function is constructed that takes behavior state parameters, environmental state parameters, and internal state parameters as inputs and viewpoint distribution parameters as outputs. ; In the formula, It is a moment The field of view distribution parameter vector includes the field of view coverage, center position, and sensitivity distribution region characteristics; These are behavioral state parameters, determined by the behavioral classification algorithm; These are environmental state parameters that describe the characteristics of the surrounding environment; These are internal state parameters that reflect the animal's physiological and cognitive state; It is a mapping function.

[0042] In this embodiment of the invention, to further enhance the realism and intelligence of the visual field changes, a behavior-visual field mapping model is established. A neural network model is trained by collecting a large amount of behavioral data (videos, IMU data) and corresponding eye-tracking data of animals in different environments. The input to this model is a multi-dimensional vector, including: behavioral state parameters. : Current behavior identified in real time from collar IMU data using behavioral classification algorithms (such as LSTM networks), such as standing, walking, running, looking down to forage, and looking up to be alert. Environmental state parameters Provided by a GPS module and pre-loaded electronic maps on the collar, such as open grasslands, dense bushes, and areas near water sources. Internal status parameters. Based on physiological rhythms and historical behavioral patterns, such as resting and active periods, the model infers these parameters. The model's output is a dynamically adjusted vector of field-of-view distribution parameters. This includes parameters such as eye rotation angle, visual axis angle, and scaling factors for each functional zone. For example, when the model identifies an animal in a "head-up alert" state, it outputs a field-of-view parameter that decreases the visual axis angle (focusing on distant objects) and expands the binocular fusion zone. This drives the binocular field-of-view parameterization model to make corresponding adjustments, resulting in a reconstructed biomimetic field of view that better reflects the animal's actual biological and behavioral patterns. This provides strong technical support for quantitative analysis of animal behavior, intelligent grazing supervision, and wildlife conservation.

[0043] Although specific embodiments of the invention have been described in detail with reference to the accompanying drawings, this should not be construed as limiting the scope of protection of this patent. Various modifications and variations that can be made by a person skilled in the art without inventive effort within the scope described in the claims still fall within the scope of protection of this patent.

Claims

1. A biomimetic field-of-view construction method based on animal physiological skeletal structure characteristics, characterized in that, include: Establish a coordinate system, which includes a geodetic coordinate system, a neck-based coordinate system, a neck-cervical vertebral joint coordinate system, and a left and right orbital coordinate system; Obtain physiological skeletal structure parameters of the animal's head and neck, and establish a kinematic model based on the parameters. The kinematic model is described by kinematic equations from the neck base coordinate system to the left and right orbit coordinate systems, and the kinematic equations pass through the coordinate systems of each neck vertebra joint in sequence. Obtain the physiological structure of the eyeball, and construct a binocular visual field parameterization model based on the kinematic model; Obtain the pose transformation matrix of the smart collar worn around the animal's neck, and solve the rotation angle of each cervical vertebra joint using inverse kinematics. Substituting the rotation angles of each cervical vertebra joint obtained by the solution into the kinematic equation, the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system is obtained; the pose transformation matrix of the left and right orbital coordinate systems in the neck base coordinate system is transformed to the geodetic coordinate system to obtain the pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system. The pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system is fused with the binocular visual field parameterization model to reconstruct the animal's biomimetic visual field in three-dimensional space.

2. The method according to claim 1, characterized in that, The head and neck skeletal parameters include cervical vertebral segment dimensions and articular surface angle range of motion, and the kinematic equations are constructed in the following manner: Obtain the cervical vertebral segment dimensions and articular surface angle range of motion; Analyze the characteristics of neck-head motion degrees of freedom to determine the vertical and horizontal rotation angles of each cervical vertebrae. A kinematic equation is constructed from the neck coordinate system to the left and right orbital coordinate systems. This kinematic equation describes the motion relationships between the cervical vertebrae through a homogeneous transformation matrix. The kinematic equation is expressed as follows: ； In the formula, This represents the homogeneous pose transformation matrix from the neck base coordinate system to the left and right eye socket coordinate systems; Indicates multiplication; From the first The cervical vertebrae to the first The homogeneous pose transformation matrix of each cervical vertebra represents the motion relationship between adjacent cervical vertebral segments. It is the first The vertical rotation angle of each cervical vertebra; It is the first The horizontal rotation angle of the cervical vertebrae; From the first A fixed transformation matrix from the cervical spine to the left and right eyes.

3. The method according to claim 2, characterized in that, The homogeneous transformation matrix between cervical vertebrae segments is: ； In the formula, It is the first The length of each cervical vertebral segment.

4. The method according to claim 1, characterized in that, The ocular physiological structure includes the eyeball radius, visual axis angle, and eyeball rotation range. The acquisition of the ocular physiological structure, based on the kinematic model, involves constructing a binocular visual field parameterized model, including: Based on the kinematic model, the initial pose of the left and right orbital coordinate systems in the neck base coordinate system is determined. The position of the eyeball center in the orbital coordinate system is determined based on the eyeball radius, and the vertex of the field cone is set to be located at the eyeball center; The relative direction of the visual axes of the left and right eyes is determined by the angle between the visual axes; Based on the range of eyeball rotation, the boundary angles of the field cones for the left and right eyes are determined respectively, and a monocular field cone constrained by the boundary angles with the visual axis as the central axis is constructed in the orbital coordinate system. Transform the left and right monocular field-of-view cones to the same coordinate system, and obtain the binocular overlap region by finding the intersection through spatial geometry. Based on overlapping and non-overlapping regions, a parameterized model containing field-of-view boundary curves and sensitivity distribution information is generated.

5. The method according to claim 4, characterized in that, Based on the field of view boundary curve and sensitivity distribution information, the field of view is divided into multiple functional zones; wherein, the functional zones include the central binocular fusion zone, the binocular fusion transition zone, the unilateral eye field of view zone, and the peripheral low-sensitivity zone.

6. The method according to claim 2, characterized in that, The process of obtaining the pose transformation matrix of the smart collar worn around the animal's neck and solving for the rotation angles of each cervical vertebra joint using inverse kinematics includes: Based on the pose transformation matrix of the smart collar, the cervical spine motion parameter vector is solved using a nonlinear optimization method: ； In the formula, It is a cervical spine motion parameter vector, which includes the rotation angles of all cervical vertebrae; It is a pose transformation matrix obtained by monitoring the smart collar worn around the neck; It is the calibration offset matrix; It is the Frobenius norm.

7. The method according to claim 1, characterized in that, The step of transforming the pose transformation matrix of the left and right orbit coordinate systems in the neck base coordinate system to the geodetic coordinate system, to obtain the pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system, includes: Obtain the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system. ; The pose transformation matrix of the left and right eye socket coordinate systems in the neck base coordinate system. Multiply the pose transformation matrix of the neck base coordinate system in the geodetic coordinate system The pose transformation matrix of the left and right orbital coordinate systems in the geodetic coordinate system is obtained. : 。 8. The method according to claim 1, characterized in that, The process of fusing the pose transformation matrices of the left and right orbital coordinate systems in the geodetic coordinate system with the binocular visual field parameterization model to reconstruct the animal's biomimetic visual field in three-dimensional space includes: Based on the left and right eye field conic geometric parameters and functional partition boundaries defined in the binocular field parametric model, a field geometry fixed to the left and right eye sockets is established in virtual three-dimensional space. Obtain the pose transformation matrix of the left and right orbit coordinate systems in the geodetic coordinate system in real time; The field-of-view geometry is transformed to the geodetic coordinate system according to the corresponding pose transformation matrix to obtain the position, orientation and coverage area of ​​the left and right fields of view in the geodetic coordinate system. The transformed left and right fields of view are spatially superimposed to generate a complete dynamic bionic field of view that includes the binocular overlap area and the monocular area.

9. The method according to claim 1, characterized in that, Also includes: A behavior-viewpoint mapping model is established. By analyzing the variation patterns of viewpoint parameters under different behavior states, a mapping function is constructed that takes behavior state parameters, environmental state parameters, and internal state parameters as inputs and viewpoint distribution parameters as outputs. ； In the formula, It is a moment The field of view distribution parameter vector includes the field of view coverage, center position, and sensitivity distribution region characteristics; These are behavioral state parameters, determined by the behavioral classification algorithm; These are environmental state parameters that describe the characteristics of the surrounding environment; These are internal state parameters that reflect the animal's physiological and cognitive state; It is a mapping function.

Images