Dynamic bird flock observation method and system based on neural hybrid camera

Abstract

The invention discloses a dynamic bird flock observation method and system based on a neural hybrid camera. The method comprises the following steps: step 1, obtaining a center view and a surrounding view of a bird flock; step 2, according to the # imgabs0 # imgabs1 #, firstly using the center view to train a neural center view field CItheta of the # imgabs2 #, and then using the surrounding view to train a parallax field # imgabs4 # of the # imgabs3 #; and step 3, according to the biplane light field coordinates under the new view angle, using the neural light field network model # imgabs5 # to output a panoramic view of the dynamic bird flock under the new view angle. According to the invention, the image of any light field visual angle can be rendered only through the image data collected by the hybrid camera system, and the light field color and angle information of the hybrid camera data can be fully mined.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision and computer graphics, and in particular to a method and system for observing a dynamic bird flock based on a neural hybrid camera. Background Art

[0002] In the observation of dynamic bird flocks, there are many challenges in achieving a large field of view, high resolution, and three-dimensional imaging. First, bird flocks usually perform complex flight activities in a vast space, requiring the observation system to have a wide field of view to capture the dynamic changes of the entire flock. It is difficult for traditional imaging equipment to strike a balance between a large field of view and high resolution. A field of view that is too small may not be able to fully cover the activities of the bird flock, while an expanded field of view usually leads to a decrease in resolution. In addition, the flight trajectories of bird flocks are complex, often accompanied by high deformation and rapid movement, which makes it more difficult for the observation system to perform three-dimensional imaging. High-quality three-dimensional imaging requires both high spatial and temporal resolution to ensure that the fine movement trajectories of the bird flocks and their distribution in three-dimensional space are captured, but this requirement often brings a lot of data processing and storage pressure.

[0003] The neural hybrid light field camera has significant advantages in the observation of dynamic bird flocks. First, it can greatly improve the imaging field of view, so that bird flock activities in a larger range can be accurately captured, while maintaining high-resolution light field output to ensure that the observation system can obtain detailed information about the bird flock. In addition, the neural hybrid light field camera can store light field data in a neural network, greatly reducing the amount of data required for traditional light field storage. This storage method not only effectively alleviates the data processing and storage bottlenecks generated by high-resolution and large-field-of-view observations, but also improves the efficiency of data transmission and processing, allowing the observation of dynamic bird flocks to be efficiently realized under large-scale and high-precision requirements. Summary of the invention

[0004] The purpose of the present invention is to provide a dynamic bird flock observation method and system based on a neural hybrid camera to meet the requirements of dynamic bird flock observation to achieve large field of view, high resolution and three-dimensional imaging.

[0005] To achieve the above object, the present invention provides a dynamic bird flock observation method based on a neural hybrid camera, which comprises:

[0006] Step 1: Collect bird flock image data of different resolutions through a hybrid camera system. The hybrid camera system includes a camera arranged at the central viewpoint (u0, v0) to obtain the central view. The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) to obtain surrounding views.

[0007] Step 2, using the center view from step 1 Neural Light Field Network Model The Neurocentric View Field CI θ (x, y) for training, θ represents the neural center view field CI θ (x,y) network parameter set, using the surrounding view obtained in step 1 Neural Light Field Network Model Neural disparity field For training, φ represents the neural disparity field A set of network parameters; among them, the neural light field network model Ways to obtain include:

[0008] The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x,y) and the neural disparity field Represent the center view function and the disparity function d u,v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1)

[0009] In the formula, Represents the surrounding views of the light field viewpoint (u, v) obtained by the central view and the disparity map relative to the central viewpoint physical process;

[0010] Step 3: Based on the biplane light field coordinates (u, v, x, y) under the new perspective, use the neural light field network model from step 2 The biplane light field coordinates (u, v, x, y) are mapped to corresponding color values ​​to obtain the image information of the dynamic bird flock at that location, and then output a panoramic view of the dynamic bird flock from a new perspective.

[0011] Furthermore, the neural center view field CI of step 2 θ The training method of (x, y) specifically includes: using the loss function described by formula (5) to obtain the neural center view field CI θ (x,y):

[0012]

[0013] In the formula, CI θ (x,y) and The reconstruction loss, Represents the central view with a resolution of αH×αW π in a biplane light field xy The coordinate set of the plane, || means taking the absolute value.

[0014] Furthermore, the neural disparity field of step 2 The training method specifically includes: using the loss function described by formula (9) Obtaining Neural Disparity Fields

[0015]

[0016] Among them, λ1,λ2,λ3 are hyperparameters, is the structure-aware smoothness loss described by equation (6), is the structural similarity index loss described by formula (7), is the surrounding view described by equation (8) The predicted value of and surrounding views reconstruction losses;

[0017]

[0018] In the formula, x S Represented as the surrounding view with resolution H×W π in the two-plane light field uv Plane, π xy The coordinate set of the plane, represents the gradient along the horizontal and vertical directions, ⊙ represents the Hadamard product, and η represents the adjustment parameter for controlling the strength of edge sensitivity. It represents the disparity map of the light field viewpoint (u, v) relative to the central viewpoint (u0, v0), || represents the absolute value, and SSIM is an indicator used to measure the similarity of two images.

[0019] Furthermore, λ1=1, λ2=0.1, λ3=0.01.

[0020] The present invention also provides a dynamic bird flock observation system based on a neural hybrid camera, which comprises:

[0021] A flock image acquisition unit is used to acquire flock image data of different resolutions through a hybrid camera system. The hybrid camera system includes a camera arranged at a central viewpoint (u0, v0) to obtain a central view. The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) to obtain surrounding views.

[0022] Neural Light Field Network model creation unit for utilizing the center view of the bird flock image acquisition unit Neural Light Field Network Model The Neurocentric View Field CI θ (x, y) for training, θ represents the neural center view field CI θ (x,y) network parameter set, using the surrounding view of the bird flock image acquisition unit Neural Light Field Network Model Neural disparity field For training, φ represents the neural disparity field A set of network parameters; among them, the neural light field network model Ways to obtain include:

[0023] The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x,y) and the neural disparity field Represent the center view function and the disparity function d u,v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1)

[0024] In the formula, Represents the surrounding views of the light field viewpoint (u, v) obtained by the central view and the disparity map relative to the central viewpoint physical process;

[0025] The scene reconstruction unit is used to use the neural light field network model in step 2 according to the biplane light field coordinates (u, v, x, y) under the new perspective. The biplane light field coordinates (u, v, x, y) are mapped to corresponding color values ​​to obtain the image information of the dynamic bird flock at that location, and then output a panoramic view of the dynamic bird flock from a new perspective.

[0026] Furthermore, the neural center view field CI θ The training method of (x, y) specifically includes: using the loss function described by formula (5) to obtain the neural center view field CI θ (x,y):

[0027]

[0028] In the formula, CI θ (x,y) and The reconstruction loss, Represents the central view with a resolution of αH×αW π in a biplane light field xy The coordinate set of the plane, || means taking the absolute value.

[0029] Furthermore, the neural disparity field The training method specifically includes: using the loss function described by formula (9) Obtaining Neural Disparity Fields

[0030]

[0031] Among them, λ1,λ2,λ3 are hyperparameters, is the structure-aware smoothness loss described by equation (6), is the structural similarity index loss described by formula (7), is the surrounding view described by equation (8) The predicted value of and surrounding views reconstruction losses;

[0032]

[0033] In the formula, x S Represented as the surrounding view with resolution H×W π in the two-plane light field uv Plane, π xy The coordinate set of the plane, represents the gradient along the horizontal and vertical directions, ⊙ represents the Hadamard product, and η represents the adjustment parameter for controlling the strength of edge sensitivity. It represents the disparity map of the light field viewpoint (u, v) relative to the central viewpoint (u0, v0), || represents the absolute value, and SSIM is an indicator used to measure the similarity of two images.

[0034] Furthermore, λ1=1, λ2=0.1, λ3=0.01.

[0035] The present invention can achieve the goal of not using a large light field camera array, but using a hybrid camera system to specifically design the arrangement and number of cameras, synchronously capturing bird dynamic data, rendering images of any light field perspective only through image data collected by the hybrid camera system, and being able to fully mine the light field color and angle information of the hybrid camera data, which helps to reconstruct the flight trajectory and posture of the bird flock under different perspectives and lighting conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1It is a flow chart of a dynamic bird flock observation method based on a neural hybrid camera provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In the drawings, the same or similar reference numerals are used to represent the same or similar elements or elements with the same or similar functions. The embodiments of the present invention are described in detail below in conjunction with the drawings.

[0038] In the description of the present invention, the terms "center", "longitudinal", "lateral", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the scope of protection of the present invention.

[0039] Terminology explanation: Neural Fields is a type of technology that implicitly represents continuous fields through neural networks, and is usually used to represent physical quantities or properties in space. It takes continuous spatial information (such as coordinates or time) as input to the neural network, and outputs some properties in the space through the network, such as color, density, lighting, geometry, etc. Neural fields can be widely used in 3D reconstruction, scene representation, physical simulation and other fields.

[0040] like Figure 1 As shown, the dynamic bird flock observation method based on the neural hybrid camera provided by the embodiment of the present invention includes:

[0041] Step 1: collect bird flock image data of different resolutions through a hybrid camera system, and obtain mixed light field data after preprocessing the bird flock image data.

[0042] The hybrid camera system includes a camera arranged at the central viewpoint (u0, v0) for collecting high-resolution images of the flock of birds, hereinafter referred to as the central view The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) for collecting low-resolution images of the flock of birds, hereinafter referred to as the surrounding viewpoints. By presetting the relative position relationship between the camera at the central viewpoint and the cameras at the surrounding positions of the central viewpoint, the position information of the flock of birds relative to the camera can be obtained, and then the position information of the flock of birds can be obtained. Therefore, the image of the flock of birds contains both the position information and the posture information of the flock of birds.

[0043] The preprocessing method may include but is not limited to using the imresize method of Matlab to perform image scaling processing.

[0044] Step 2: Since the center view captured by the hybrid camera The central view obtained in step 1 is used to estimate the color of the scene. Neural Light Field Network Model The Neurocentric View Field CI θ (x,y) for training. Neural Center View Field CI θ The input (x,y) is a two-dimensional vector, π in the light field biplane xy The plane two-dimensional position coordinates x = (x, y), the two-dimensional position coordinates are sent to the MLP for training, such as Output the color c of the sampling point in four-dimensional space.

[0045] θ represents the neural center view field CI θ The network parameter set of (x, y), θ, can be set as, but not limited to: Neural Center View Field CI θ The training process uses a 6-layer MLP (Multilayer Perceptron in English and fully connected network layer in Chinese), uses a sinusoidal activation function, and the size of the neurons in the middle layer is 256. The position encoding of the image space coordinates is set to 8, and the Adam optimizer is used in the training process. The initial learning rate is 1×10 -2 , the final learning rate is 1×10 -5 , it decreases by 0.1 times every 4000 rounds.

[0046] Using the surrounding view obtained in step 1 Neural Light Field Network Model Neural disparity field Training. Neural Disparity Field The input is a four-dimensional vector, the four-dimensional position coordinates of the light field biplane r = (u, v, x, y), and the four-dimensional vector is sent to the MLP for training, such as Output the disparity d(u,v,x,y) of the sampling point in four-dimensional space (u,v,x,y).

[0047] φ represents the neural disparity field The network parameter set, φ, can be set to, for example but not limited to, the following: Neural disparity field It is a 5-layer MLP with Relu activation function. The neuron size of the middle layer is 256. The light field angle and spatial position encoding are set to 6 and 8 respectively. The training process uses Adam optimizer with a learning rate of 1×10 -2 .

[0048] Implicit neural representation uses a neural network to learn a continuous function to represent the characteristics of an object or certain information in a spatial scene, such as color, density, depth, etc. In this embodiment, there are two types of implicit neural representations: using the neural center view field CI θ (x,y) Learn the central view function of the light field using neural disparity field Learning the disparity function of the light field. The dynamic adjustment of the parameters and spectral characteristics required for different scenes means that the resolution and spectral characteristics of the light field data collected in different scenes are different. The neural center view field CI constructed in this paper θ (x,y) and the neural disparity field It is necessary to adjust the network parameters such as the number of MLP layers N, the number of neurons W, and the selection of activation functions, as well as the dimension L of position encoding. When position encoding the light field biplane coordinates, the hyperparameters are adjusted according to the frequency characteristics of the light field space and angle. The activation function of the neural disparity field is the ReLU activation function, and the neural central view field is the sine activation function.

[0049] Among them, the neural light field network model Ways to obtain include:

[0050] The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x,y) and the neural disparity field Represent the center view function and the disparity function d u,v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1)

[0051] In the formula, Represents the surrounding views of the light field viewpoint (u, v) obtained by the central view and the disparity map relative to the central viewpoint The physical process, θ represents the neural center view field CI θ (x,y) is a set of network parameters, φ represents the neural disparity field A collection of network parameters.

[0052] In one embodiment, the neural center view field CI θ The training method of (x,y) specifically includes:

[0053] Use the loss function used to measure the gap between the model prediction value and the true value to perform gradient backpropagation to obtain the neural center view field CI θ(x, y). For example: Using the loss function described in formula (5), we can obtain the neural center view field CI θ (x,y):

[0054]

[0055] In the formula, CI θ (x,y) and The reconstruction loss, Represents the central view with a resolution of αH×αW π in a biplane light field xy The coordinate set of the plane, || represents the absolute value.

[0056] In one embodiment, the neural disparity field The training method specifically includes: using the loss function described by formula (9) Obtaining Neural Disparity Fields

[0057]

[0058] Among them, λ1, λ2, λ3 are hyperparameters, and their specific values ​​are adjusted according to the specific task, for example, λ1=1, λ2=0.1, λ3=0.01, is the structure-aware smoothness loss described by equation (6), is the structural similarity index SSIM (full name Structural Similarity Index) loss described by formula (7), For neural disparity fields In the surrounding view and Neurocenter View Field CI θ Perform region matching between continuous data of (x,y). is the surrounding view described by equation (8) The predicted value of and surrounding views reconstruction losses;

[0059]

[0060] In the formula, x S Represented as the surrounding view with resolution H×W π in the two-plane light field uv Plane, π xy The coordinate set of the plane, represents the gradient along the horizontal and vertical directions, ⊙ represents the Hadamard product, and η represents the adjustment parameter for controlling the strength of edge sensitivity. Represents the disparity map of the light field viewpoint (u, v) relative to the central viewpoint (u0, v0), || represents the absolute value, From step 2 SSIM is an indicator used to measure the similarity between two images, and is particularly suitable for tasks related to image quality assessment and visual perception. SSIM takes into account the sensitivity of the human eye to image structure, brightness, and contrast, so it is more in line with human visual characteristics in image quality evaluation.

[0061] Step 3: Based on the biplane light field coordinates (u, v, x, y) under the new perspective, these coordinates represent the π of the virtual camera in the light field. uv Planes and π xy The sampling position of the plane, using the neural light field network model in step 2 The biplane light field coordinates (u, v, x, y) are mapped to corresponding color values ​​to obtain the image information of the dynamic bird flock at that location. For the color value of each sampling point The color data of each position is combined into a complete image. In this way, the final output is a panoramic view of the dynamic bird flock from a new perspective, which can accurately restore the flight posture and spatial relationship of the bird flock.

[0062] The embodiment of the present invention also provides a dynamic bird flock observation system based on a neural hybrid camera, which includes a bird flock image acquisition unit, a neural light field network model creation unit and a scene reconstruction unit:

[0063] The bird flock image acquisition unit is used to collect bird flock image data of different resolutions through a hybrid camera system. The hybrid camera system includes a camera arranged at the central viewpoint (u0, v0) to obtain the central view. The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) to obtain surrounding views.

[0064] Neural Light Field Network Model Creation Unit is used to utilize the center view of the bird flock image acquisition unit Neural Light Field Network Model The Neurocentric View Field CI θ (x, y) for training, θ represents the neural center view field CI θ (x,y) network parameter set, using the surrounding view of the bird flock image acquisition unit Neural Light Field Network Model Neural disparity field For training, φ represents the neural disparity field A set of network parameters; among them, the neural light field network model Ways to obtain include:

[0065] The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x,y) and the neural disparity field Represent the center view function and the disparity function d u,v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1)

[0066] The scene reconstruction unit is used to utilize the neural light field network model Get the color value of any coordinate in the biplane light field and compare the color value with the neural light field network model The corresponding angle information contained is combined to generate a new perspective and flight posture diagram of the dynamic bird flock.

[0067] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Those skilled in the art should understand that the technical solutions described in the above embodiments may be modified, or some of the technical features thereof may be replaced by equivalents; these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic bird flock observation based on a neural hybrid camera, characterized in that: include: Step 1: Collect bird flock image data of different resolutions through a hybrid camera system. The hybrid camera system includes a camera arranged at the central viewpoint (u0, v0) to obtain the central view. The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) for acquiring surrounding views. Step 2, using the center view from step 1 Neural Light Field Network Model The Neurocentric View Field CI θ (x, y) for training, θ represents the neural center view field CI θ The network parameter set of (x, y) uses the surrounding view obtained in step 1 Neural Light Field Network Model Neural disparity field For training, φ represents the neural disparity field A set of network parameters; among them, the neural light field network model Ways to obtain include: The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x, y) and the neural disparity field Represent the center view function and the disparity function d u，v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1) In the formula, Represents the surrounding views of the light field viewpoint (u, v) obtained by the central view and the disparity map relative to the central viewpoint physical process; Step 3: Based on the biplane light field coordinates (u, v, x, y) under the new perspective, use the neural light field network model from step 2 The biplane light field coordinates (u, v, x, y) are mapped to corresponding color values ​​to obtain the image information of the dynamic bird flock at that location, and then output a panoramic view of the dynamic bird flock at a new perspective.

2. The method for dynamic bird flock observation based on neural hybrid camera according to claim 1, characterized in that: Step 2 of the Neurocentric View Field CI θ The training method of (x, y) specifically includes: using the loss function described by formula (5) to obtain the neural center view field CI θ (x, y): In the formula, l h CI θ (x, y) and The reconstruction loss, Represents the central view with a resolution of αH×aW π in a biplane light field xy The coordinate set of the plane, || means taking the absolute value.

3. The method for dynamic bird flock observation based on a neural hybrid camera according to claim 1 or 2, characterized in that: Neural disparity field at step 2 The training method specifically includes: using the loss function l described by formula (9) l Obtaining Neural Disparity Fields l l =λ1l1+λ2l SSIM +λ3l smooth (9) Among them, λ1, λ2, λ3 are hyperparameters, l smooth is the structure-aware smoothness loss described by equation (6), l SSIM is the structural similarity index loss described by formula (7), l1 is the surrounding view described by formula (8) The predicted value of and surrounding views reconstruction losses; In the formula, x S Represented as the surrounding view with resolution H×W π in the two-plane light field uv Plane, π xy The coordinate set of the plane, represents the gradient along the horizontal and vertical directions, ⊙ represents the Hadamard product, and η represents the adjustment parameter for controlling the strength of edge sensitivity. It represents the disparity map of the light field viewpoint (u, v) relative to the central viewpoint (u0, v0), || represents the absolute value, and SSIM is an indicator used to measure the similarity of two images.

4. The method for dynamic bird flock observation based on neural hybrid camera according to claim 3, characterized in that: λ1=1, λ2=0.1, λ3=0.

01.

5. A dynamic bird flock observation system based on neural hybrid camera, characterized in that: include: A flock image acquisition unit is used to acquire flock image data of different resolutions through a hybrid camera system. The hybrid camera system includes a camera arranged at a central viewpoint (u0, v0) to obtain a central view. The hybrid camera system also includes cameras arranged around the central viewpoint (u0, v0) for acquiring surrounding views. Neural Light Field Network model creation unit for utilizing the center view of the bird flock image acquisition unit Neural Light Field Network Model The Neurocentric View Field CI θ (x, y) for training, θ represents the neural center view field CI θ (x, y) network parameter set, using the surrounding view of the bird flock image acquisition unit Neural Light Field Network Model Neural disparity field For training, φ represents the neural disparity field A set of network parameters; among them, the neural light field network model Ways to obtain include: The position coordinates (u, v, x, y) of the biplane light field are used as the input of the parallax propagation equation L(u, v, x, y) for anti-occlusion and anti-radiation described by equation (2), and the neural center view field CI is used to calculate the θ (x, y) and the neural disparity field Represent the center view function and the disparity function d u，v (x, y), as shown in equations (3) and (4), we get the neural light field network model described by equation (1) In the formula, Represents the surrounding views of the light field viewpoint (u, v) obtained by the central view and the disparity map relative to the central viewpoint physical process; A scene reconstruction unit is used to use the neural light field network model in step 2 according to the biplane light field coordinates (u, v, x, y) under the new perspective. The biplane light field coordinates (u, v, x, y) are mapped to corresponding color values ​​to obtain the image information of the dynamic bird flock at that location, and then output a panoramic view of the dynamic bird flock at a new perspective.

6. The dynamic bird flock observation system based on neural hybrid camera according to claim 5, characterized in that: NeuroCenter View Field CI θ The training method of (x, y) specifically includes: using the loss function described by formula (5) to obtain the neural center view field CI θ (x, y): In the formula, l h CI θ (x, y) and The reconstruction loss, Represents the central view with a resolution of αH×αW π in a biplane light field xy The coordinate set of the plane, || means taking the absolute value.

7. The dynamic bird flock observation system based on neural hybrid camera according to claim 5 or 6, characterized in that: Neural disparity field The training method specifically includes: using the loss function l described by formula (9) l Obtaining Neural Disparity Fields l l =λ1l1+λ2l SSIM +λ3l smooth (9) Among them, λ1, λ2, λ3 are hyperparameters, l smooth is the structure-aware smoothness loss described by equation (6), l SSIM is the structural similarity index loss described by formula (7), l1 is the surrounding view described by formula (8) The predicted value of and surrounding views reconstruction losses; In the formula, x S Represented as the surrounding view with resolution H×W π in the two-plane light field uv Plane, π xy The coordinate set of the plane, represents the gradient along the horizontal and vertical directions, ⊙ represents the Hadamard product, and η represents the adjustment parameter for controlling the strength of edge sensitivity. It represents the disparity map of the light field viewpoint (u, v) relative to the central viewpoint (u0, v0), || represents the absolute value, and SSIM is an indicator used to measure the similarity of two images.

8. The method for dynamic bird flock observation based on neural hybrid camera according to claim 7, characterized in that: λ1=1, λ2=0.1, λ3=0.01.

Images