Underwater environment three-dimensional reconstruction method

Abstract

The invention relates to an underwater environment three-dimensional reconstruction method. The method comprises the following steps: projecting a multi-band structured light pattern in an underwater environment, and capturing an image by using a multispectral camera; fusing images of different wavebands through a spectrum matching algorithm, and extracting depth information to generate three-dimensional reconstruction basic data; a depth error caused by underwater scattering and absorption is corrected by using a depth compensation technology; generating an environment characteristic model by using real-time monitoring environment parameters including illumination, turbidity and temperature of an underwater sensor; adaptively adjusting image acquisition and processing parameters including exposure, contrast and color balance according to the environment characteristic model; detecting a distorted area caused by environment change in the image through an intelligent correction algorithm, and executing local correction and overall optimization; generating a multi-scale image pyramid, and decomposing the structure information of the three-dimensional image layer by layer from coarse to fine; and through an adaptive detail enhancement technology, carrying out refinement processing on the key area to generate an underwater three-dimensional image.

Description

Technical Field

[0001] The invention relates to the field of underwater environment modeling, and in particular to a method for three-dimensional reconstruction of an underwater environment. Background Art

[0002] Although the current methods for 3D reconstruction of underwater environments have made certain technical progress, there are still some significant deficiencies and drawbacks in practical applications, which largely limit the application effect of 3D reconstruction technology in underwater environments. First, the complexity of the underwater environment poses huge challenges to traditional 3D reconstruction methods. Underwater optical imaging is affected by light attenuation, scattering and absorption, and these factors become more complicated with changes in depth, turbidity and ambient light conditions. Traditional methods usually rely on single-band optical imaging or sonar imaging technology, which perform poorly in highly turbid waters and are easily affected by light attenuation and enhanced scattering, resulting in inaccurate or lost depth information. At the same time, traditional 3D reconstruction methods often assume that light propagates uniformly in water, but in practical applications, the heterogeneity and dynamic changes of water bodies make this assumption invalid, resulting in large errors in the reconstruction results.

[0003] Secondly, most of the current 3D reconstruction methods rely on static or offline environmental parameter monitoring and modeling, and lack the ability to dynamically respond to real-time data. This method often cannot quickly adjust the modeling strategy when dealing with rapidly changing conditions such as illumination, turbidity and temperature in underwater environments, resulting in the generated 3D model being easily distorted or unstable. Traditional technologies often use fixed correction parameters in underwater environments, which are difficult to adapt to complex and changing underwater conditions, especially when there are significant differences in different depths and regional environments. The generalization ability of the model is poor and the performance is unstable. The application of fixed parameters will cause the performance of the model to be significantly different under different water conditions. This inconsistency seriously affects the accuracy and reliability of 3D reconstruction. In addition, traditional methods usually use a single-level modeling and optimization strategy, ignoring the changes in environmental features at different scales, making it difficult to capture subtle local features, resulting in insufficient global consistency and local detail retention of the model. Especially in complex underwater environments, local feature changes caused by factors such as water flow and particulate matter will have a significant impact on the effect of 3D reconstruction, and a single-level modeling method is difficult to cope with such diversity and variability. In terms of image processing and feature extraction, traditional methods mainly rely on manually designed feature extraction methods, which usually extract simple low-level features (such as edges, textures, etc.) and are difficult to effectively capture complex nonlinear distortion features. This simple data fusion method often performs poorly in dealing with complex environmental changes, and it is difficult to achieve in-depth analysis and optimization of multi-source data, resulting in inaccurate and unstable reconstruction results. In terms of intelligence, traditional underwater 3D reconstruction technology still has significant shortcomings.

[0004] At present, most methods mainly rely on preset rules and parameters when dealing with image distortion and environmental changes, and lack the ability of adaptive learning and optimization. In addition, most of the existing underwater 3D reconstruction methods do not fully utilize the latest progress in deep learning and artificial intelligence. Although deep learning technology has achieved remarkable results in the field of computer vision, its application in underwater environments is still limited. Summary of the invention

[0005] The purpose of the present invention is to provide a method for three-dimensional reconstruction of an underwater environment, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.

[0006] The present invention solves the above-mentioned technical problems by adopting the following technical solutions: a method for three-dimensional reconstruction of an underwater environment, S1, by projecting an optimized spectral pattern onto a target surface, then capturing the reflected light with a camera, and inferring the three-dimensional structure of the object surface through the deformation of light;

[0007] in:

[0008] S1.1, projecting multi-band structured light patterns in an underwater environment while capturing images using a multispectral camera;

[0009] S1.2, through the spectrum matching algorithm, the images of different bands are fused, the depth information is extracted and the basic data for 3D reconstruction is generated;

[0010] S1.3. Use depth compensation technology to correct the depth error caused by underwater scattering and absorption;

[0011] S2, dynamically adjust the image correction strategy by real-time monitoring of underwater environmental parameters;

[0012] in:

[0013] S2.1. Use underwater sensors to monitor environmental parameters including light, turbidity, and temperature in real time and generate environmental characteristic models;

[0014] S2.2, adaptively adjust image acquisition and processing parameters including exposure, contrast and color balance according to the environmental feature model;

[0015] S2.3, through intelligent correction algorithm, detect the distorted areas in the image caused by environmental changes, and perform local correction and overall optimization;

[0016] S3, by constructing a multi-level image pyramid model, first capture the overall structure at a large scale, and then gradually enhance the texture details at a small scale;

[0017] in:

[0018] S3.1. Generate a multi-scale image pyramid to decompose the structural information of the three-dimensional image layer by layer from coarse to fine;

[0019] S3.2, using a generative adversarial network to enhance details and restore textures at each layer;

[0020] S3.3. Through adaptive detail enhancement technology, the key areas are refined to generate underwater three-dimensional images.

[0021] Furthermore, the spectrum matching algorithm construction process includes:

[0022] S1. Using multispectral imaging technology, multiple spectral information of visible light, infrared light and ultraviolet light is integrated into a multidimensional spectral feature matrix. The depth information of each pixel is calculated through the spectral response difference of each band. The calculation formula is expressed as:

[0023]

[0024] Where D(x,y) represents the depth value at the pixel point (x,y), λ is the spectral wavelength, S i (λ) is the spectral response of the i-th band, α i and β i (λ) is the coefficient and absorption characteristics used to weigh the contributions of different bands;

[0025] S2. Adjust the parameters of the spectral matching algorithm according to the real-time changes of the underwater environment; dynamically update the matching function to ensure the accuracy of the image by monitoring the illumination and turbidity environmental parameters; the correction process is achieved through the adaptive adjustment function:

[0026]

[0027] Among them, Φ(t) is the spectral matching function after dynamic adjustment, γ(t) represents the influence of environmental parameters on the correction function, κ is the adjustment frequency factor, and Γ(t) is the composite function of ambient light intensity and turbidity;

[0028] S3. Using the nonlinear fusion model, the spectral information of different bands is integrated into a multi-dimensional feature space, and multi-level feature mapping and deep information extraction are performed through a deep learning network. The process of nonlinear fusion and deep extraction is expressed by the following formula:

[0029]

[0030] Among them, Z(x,y) represents the depth information after nonlinear fusion at the pixel point (x,y), Λ is the spectral range, δ j and ω j is the weight coefficient and frequency factor of different bands, T j(λ) is the nonlinear spectral response function of the jth band.

[0031] Furthermore, the depth compensation technology adopts a solution including:

[0032] S1. By establishing an underwater optical transmission model, the scattering path and absorption coefficient of light during underwater propagation are analyzed, and the nonlinear effects of scattering and absorption on depth measurement are dynamically calculated; simulation is performed using the path integral formula, which is expressed as follows:

[0033]

[0034] Among them, I(x,y,z) represents the light intensity at the point (x,y,z), which describes the energy of light propagating in an underwater environment; λ is the wavelength of light, and α(λ) is the absorption coefficient, which represents the degree of attenuation of light at a specific wavelength;

[0035] The spectral response function η(λ) of the scattering path describes the scattering behavior of light of different wavelengths when passing through water, θ(s) represents the angular change of the scattering path, and the integral path L represents the path length of the light propagating in the water;

[0036] S2. Dynamically adjust the parameters of the correction model according to the underwater environmental parameters monitored in real time; through a multi-level correction mechanism, linear correction of the light scattering coefficient is used in shallow areas, and depth compensation based on nonlinear scattering paths is used in deep and high turbidity areas; this is achieved through the following formula:

[0037]

[0038] Where D(z) represents the corrected depth value at depth z, which is dynamically calculated through a multi-level correction mechanism; in the formula D(z) represents the shallow linear correction, where γ n is the correction factor, β n is the depth-dependent attenuation factor; and Then the nonlinear correction term δ is introduced n (z), used to handle nonlinear depth compensation in water environments;

[0039] S3, through the parallel processing architecture, scatter compensation and depth reconstruction are performed in real time, and the compensation parameters are dynamically adjusted using the depth error feedback mechanism; the compensation and reconstruction synchronization process is achieved through the following formula:

[0040]

[0041] Among them, E(t) represents the depth error after real-time compensation, reflecting the error correction effect of the entire compensation and reconstruction process; ξ(t) in the formula E(t) is the scattering compensation coefficient, which is used to adjust the compensation intensity in real time, and ω is the synchronization frequency factor, which determines the coordination frequency between compensation and reconstruction; To reconstruct the feedback adjustment term ζ(t), the compensation effect is further optimized by performing differential operation on the real-time feedback during the reconstruction process.

[0042] Furthermore, the process of generating the environmental feature model includes:

[0043] S1. First, in the underwater environment, real-time monitoring of multi-dimensional parameters such as light, turbidity, and temperature is performed. Dynamic environmental feature extraction technology is used to perform nonlinear analysis and fusion of multi-source data. Based on real-time data, an adaptive algorithm is used to build a preliminary model of environmental features. The formula is dynamically adjusted to adapt to environmental changes during the 3D reconstruction process:

[0044]

[0045] Among them, M(t) represents the dynamic response parameter of the environmental characteristic model at time t, α(t) represents the weight coefficient of different environmental parameters including light, turbidity, and temperature, which is dynamically adjusted according to real-time monitoring data, ω is the frequency factor of the model response, reflecting the periodicity of environmental changes, and β(t) is the characteristic function that describes the nonlinear changes of environmental parameters;

[0046] S2. Then, local models are built for illumination, turbidity, and temperature to capture the characteristics of different depths and regions. The local models are integrated into a global environmental feature model and dynamically adjusted through a multi-level optimization algorithm. The optimization process is as follows:

[0047]

[0048] Among them, G(x,y,z) is the global environment feature model at the spatial position (x,y,z), γ n is a local environmental characteristic parameter, which indicates the initial value of environmental characteristics in a specific area or depth, δ n is the depth-related environmental impact factor, describing the changing trend of environmental parameters with depth, ∈ n (z) is a nonlinear optimization term used to further correct the accuracy of the model;

[0049] S3. Finally, the data from different sensors are deeply integrated through the intelligent sensor data fusion mechanism, and future environmental changes are predicted through time series analysis. The core formula for data fusion and prediction is:

[0050]

[0051] Among them, P(t) represents the environmental change prediction parameter at time t, λ(t) is the fusion weight of different sensor data, which determines the influence of each sensor in the prediction result, κ is the frequency factor of the time series, which reflects the periodicity and trend of environmental changes, and θ(t) is the prediction correction term, which is used to correct the prediction result.

[0052] Furthermore, the construction method steps adopted by the intelligent correction algorithm include:

[0053] S1. Identify local distortion areas caused by environmental changes through multi-dimensional matching of environmental parameters and image features; and distinguish and locate distortion areas based on an intelligent distortion detection model;

[0054] S2, use nonlinear optimization method to gradually correct the distorted area; at the same time, combine multi-scale optimization technology to perform layered processing on the entire image;

[0055] S3. Use the convolutional neural network (CNN) to extract and analyze the features of the distorted area, and then use the generative adversarial network (GAN) to achieve high-quality image reconstruction.

[0056] Furthermore, the construction process of the local distortion area identification method includes:

[0057] S1. By comprehensively matching the environmental parameters of illumination, turbidity, and temperature with the image features of brightness, texture, and edge, a high-dimensional feature space is formed to identify the local distortion areas caused by environmental changes; by analyzing the nonlinear relationship between environmental parameters and image features, a multi-dimensional matching matrix is ​​generated, which is mathematically expressed as follows:

[0058]

[0059] Among them, F(x,y,t) represents the matching function value at the spatial position (x,y) and time t, representing the comprehensive influence between environmental parameters and image features; parameter α i (t) is the weight of the environmental parameters including illumination, turbidity, and temperature, which controls the contribution of each environmental factor to image distortion. i (x, y, t) is the nonlinear response function of the image feature, which is used to capture the impact of environmental changes on image quality; the correction term β i (t) is used to adjust the model and improve the accuracy of distortion recognition;

[0060] S2. Based on the deep learning algorithm, the intelligent distortion detection model learns different types of distortion features through training of environmental data and image samples, and establishes a multi-level distortion recognition system; the adaptive learning process is expressed by the following formula:

[0061]

[0062] Among them, S(t) represents the distortion detection result at time t, reflecting the comprehensive judgment of the current environment and image features; parameter γ j (t) is the weight factor of the distortion feature, which is used to measure the importance of various distortion features in detection. j (t) is the corresponding characteristic response function, which describes the dynamic behavior of the distortion feature; the correction term δ j (t) is used to adjust the sensitivity and adaptability of the model;

[0063] S3. Finally, through the collaborative work of multi-dimensional matching and intelligent detection model, the environmental parameters and image features are analyzed and processed synchronously; and the collaborative detection mechanism combines the joint evaluation of the matching matrix and the intelligent model to perform in-depth analysis and positioning of the local distortion area; and the collaborative detection formula is as follows:

[0064]

[0065] Among them, C(t) represents the collaborative detection result at time t, which is a comprehensive evaluation of the coupling relationship between environmental parameters and image features; the parameter λ(t) is the weight of collaborative detection, which measures the interaction between environmental and image features in distortion detection; ω is the frequency factor of the time series, which is used to capture the periodic changes of environmental and image features; and the correction term θt is used to dynamically adjust the detection results.

[0066] Furthermore, the distorted area is gradually corrected by performing multiple rounds of iterative correction on the distorted area by combining dynamic feedback of environmental parameters and image features; the expression of the optimization process is as follows:

[0067]

[0068] Among them, L(t) represents the nonlinear loss function at time t, α i is the weight factor of environmental parameters and image features, Φ i (x, y, t) is the nonlinear response function of a specific distortion region, describing the impact of environmental changes on the image, and β(t) is the correction term used for optimization;

[0069] The hierarchical processing of the image is to divide the image into multiple scale levels and perform optimization processing at different scales; the expression of the multi-scale optimization process is as follows:

[0070]

[0071] Among them, G(x,y,z) represents the global optimization function at the spatial position (x,y,z), γ j is the optimization parameter at different scales, δ jis a scale-dependent attenuation factor, describing the optimization effects at different levels, ∈ j (z) is the optimization correction term, which is used to balance the global structure and local details.

[0072] Furthermore, the feature extraction and analysis is performed by introducing a convolutional neural network (CNN) to gradually extract high-level features in the image through multi-layer convolution and pooling operations, including nonlinear distortion features caused by environmental changes; the expression of feature extraction is as follows:

[0073]

[0074] Among them, F(x,y,t) represents the feature map value at spatial position (x,y) and time t, α i is the weight coefficient of different convolutional layers; function σ is the activation function, which captures the feature relationship by introducing nonlinear mapping; convolution kernel weight W i The product of the input feature g(x,y,t) represents the weighted sum of the features in the convolution operation, b i is a bias term used to adjust the distribution of features;

[0075] The generator of the generative adversarial network GAN generates high-quality images using the features extracted by CNN, while the discriminator evaluates the authenticity of the generated images; the expression of the loss function is as follows:

[0076]

[0077] Among them, L(G,D) represents the adversarial loss function between the generator G and the discriminator D; is the expectation of the real image x, and the discriminator D is optimized by judging whether the input image comes from real data or the generator; It is to evaluate the images output by the generator, which deceives the discriminator by generating images.

[0078] Beneficial effects of the present invention:

[0079] By introducing dynamic monitoring and adjustment of multi-dimensional environmental parameters (such as light, turbidity, and temperature), combined with deep learning technologies such as convolutional neural networks (CNN) and generative adversarial networks (GAN), this method can effectively deal with light scattering, absorption, and other nonlinear distortion problems in complex underwater environments, and significantly improve the accuracy of three-dimensional reconstruction.

[0080] Through multi-level optimization and dynamic feedback mechanisms, high image quality and reconstruction stability can be maintained in different underwater environments, especially in high turbidity and high light fluctuation conditions. This makes the method highly adaptable in various underwater operations and able to cope with changing environmental conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 This is a flow chart of the method for three-dimensional reconstruction of an underwater environment of the present invention.

[0082] Figure 2 A flow chart of the process of constructing the spectrum matching algorithm of the present invention.

[0083] Figure 3 This is a flow chart of the scheme for adopting the depth compensation technology of the present invention. DETAILED DESCRIPTION

[0084] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0085] Combined with Figure 1 The method for three-dimensional reconstruction of underwater environment of the present invention, in the first step, is to project the optimized spectral pattern on the target surface, and then capture the reflected light by the camera, and infer the three-dimensional structure of the object surface through the deformation of light. First, a multi-band structured light pattern is projected in the underwater environment, and the image is captured by a multi-spectral camera at the same time. This process is to project a specially designed spectral pattern in the underwater environment, which covers multiple spectral bands to ensure that sufficient information is captured under different optical conditions. The multi-spectral camera can capture the light reflected from different bands, and provide multi-dimensional data support for subsequent image fusion and depth extraction. Then, the images of different bands are fused by the spectral matching algorithm, and the depth information is extracted to generate the basic data for three-dimensional reconstruction. In this step, the system fuses the images of each band captured by the multi-spectral camera, uses the spectral matching algorithm to compare and match the features of these images, and calculates the depth information of the object based on the reflection and absorption characteristics of each band to light. The spectral matching algorithm uses the difference in spectral response under different bands to accurately calculate the depth value of each pixel point, and form preliminary three-dimensional point cloud data, which provides a basis for three-dimensional reconstruction. Finally, depth compensation technology is used to correct depth errors caused by underwater scattering and absorption. In underwater environments, light is scattered and absorbed by water during propagation, resulting in certain errors in the depth information in the captured image. To solve this problem, depth compensation technology constructs a refined optical transmission model, analyzes the propagation path of light underwater, and dynamically corrects these errors in combination with real-time monitored environmental parameters, such as the turbidity of the water and the attenuation characteristics of light. Through depth compensation, the system can eliminate or significantly reduce depth measurement errors caused by light scattering and absorption, and ultimately generate accurate 3D reconstruction data.

[0086] The second step is to dynamically adjust the image correction strategy by real-time monitoring of underwater environmental parameters. First, underwater sensors, including light, turbidity, temperature and other equipment, are used to monitor environmental parameters in real time, and an environmental feature model is generated based on these data. This model reflects the complex characteristics of the current underwater environment by integrating multiple environmental parameters, providing accurate environmental background data for subsequent image processing. Then, according to the generated environmental feature model, the image acquisition and processing parameters, including exposure, contrast and color balance, are adaptively adjusted. This adjustment process is dynamic. The system analyzes the environmental feature model to identify the impact of the current environment on image acquisition, so as to adjust the various parameters of the camera equipment in real time to adapt to the changing underwater conditions and ensure the quality and consistency of the image. Finally, the intelligent correction algorithm is used to detect the distorted areas in the image caused by environmental changes, and local correction and overall optimization are performed. The intelligent correction algorithm uses the multi-dimensional matching of the environmental feature model and the image features to accurately identify the distorted areas in the image caused by environmental factors such as light changes, turbidity fluctuations or temperature changes. For these distorted areas, the system does not just make a simple global adjustment, but applies a local correction strategy to gradually correct the distorted image parts. At the same time, in order to maintain the overall consistency and visual effect of the image, the system also performs overall optimization to ensure that the corrected image maintains a high degree of coherence and accuracy in both local details and global structure.

[0087] The third step is to build a multi-level image pyramid model, first capture the overall structure at a large scale, and then gradually enhance the texture details at a small scale. First, a multi-scale image pyramid is generated to decompose the structural information of the three-dimensional image layer by layer from coarse to fine. By decomposing the original image into a multi-level structure with different resolutions, each layer represents a different scale, from the coarsest overall structure to the most detailed texture details, a multi-level image pyramid model is formed. This process helps to take into account the global structure and local details of the image during the reconstruction process, ensuring the integrity and delicacy of the reconstruction. Then, a generative adversarial network (GAN) is used to enhance details and restore textures at each layer. GAN consists of a generator and a discriminator. Through adversarial training, the generator enhances the details and restores the texture of the image layer by layer, while the discriminator evaluates the output effect of the generator, thereby prompting the generator to continuously improve its generation ability. This process is repeated at each level, so that the image structure from coarse to fine is gradually refined, and the texture information is fully restored, ensuring that the details and texture of the image can still be clearly presented even in complex underwater environments. Finally, the key areas are refined through adaptive detail enhancement technology to generate underwater three-dimensional images. Adaptive detail enhancement technology can dynamically adjust the enhancement strategy according to the characteristics of the current image and focus on the key areas in the image. These key areas often contain important structural or texture information and are the focus of image detail expression. Through adaptive enhancement strategies, the system can accurately control the strength and range of enhancement, making the details of key areas clearer. The enhanced image not only maintains overall coherence, but also has extremely high quality in details. Finally, combined with the multi-scale structure of the pyramid model, the layer-by-layer refinement of GAN, and the fine processing of adaptive detail enhancement, high-quality underwater 3D images are generated, achieving high-precision 3D reconstruction in complex underwater environments.

[0088] Embodiment 1:

[0089] if Figure 2 As shown, in this embodiment, a 3D reconstruction is performed on an underwater structure to restore the 3D structure. In this task, due to the complex underwater environment, the scattering and absorption of light make it difficult for traditional single-band imaging technology to obtain clear depth information, so multi-spectral imaging technology is used to achieve it through the integration of multiple spectral information such as visible light, infrared and ultraviolet light.

[0090] First, a multispectral camera system is used to obtain image data in different bands. The multispectral camera can capture the reflected light of visible light (400-700nm), near infrared (700-1000nm) and ultraviolet light (200-400nm). These data are recorded by the sensor as spectral response S in different bands. iIn this example, the multispectral camera is set to capture spectral information of n=5 bands, where visible light, infrared light, and ultraviolet light correspond to five bands with wavelengths of 450 nm, 550 nm, 700 nm, 800 nm, and 300 nm, respectively.

[0091] The spectral matching algorithm used calculates the depth value of each pixel based on the following formula:

[0092]

[0093] In the formula, D(x,y) represents the depth value of the pixel point (x,y), λ is the spectral wavelength, and S i (λ) is the spectral response of the ith band. The wavelength range is set from 200nm to 1000nm, covering the ultraviolet, visible light and near infrared ranges. The coefficient α i It is used to weigh the contribution of each band to the depth information. The value range can be set between 0.1 and 1, and adjusted according to the importance of different bands. For example, in this example, the coefficient α of visible light and infrared light i They are set to 0.7 and 0.8 respectively, and the UV light is set to 0.5 because the UV light is more attenuated in water and has less impact. i (λ) is used to describe the attenuation characteristics of light underwater, and its value range is set to 0.01 to 0.1 to reflect the attenuation degree of different wavelengths in water. For example, β i (λ) can be set to 0.05 for infrared rays and 0.1 for ultraviolet rays, indicating that the attenuation of ultraviolet rays is more drastic.

[0094] Substitute these parameters and calculate a certain pixel point (x=100, y=150), setting the spectral response S1(450nm)=0.8, S2(550nm)=0.9, S3(700nm)=0.85, S4(800nm)=0.7, S5(300nm)=0.6. Calculate the depth value D(100,150):

[0095]

[0096] Through numerical integration, we can get the depth value of the pixel point to be about D(100,150)≈2.5 meters. This shows that the fusion of multi-band spectral data can accurately infer the three-dimensional form of the underwater structure. Such calculations are performed on the entire image to generate complete three-dimensional point cloud data, thereby achieving high-precision three-dimensional reconstruction of the underwater environment.

[0097] Then, in order to ensure the accuracy of the image, the parameters of the spectral matching algorithm can be adjusted according to the real-time changes in the underwater environment. By real-time monitoring of environmental parameters such as underwater illumination and turbidity, the spectral matching function is dynamically updated to ensure that the image remains clear and the depth data is accurate in a changing environment. Specifically, this dynamic correction is achieved by adaptively adjusting the function Φ(t), and its formula is:

[0098]

[0099] In the formula, Φ(t) represents the dynamically adjusted spectral matching function, which integrates the impact of real-time changes in environmental parameters on the image. The parameter γ(t) is used to represent the impact of environmental parameters on the correction function. These environmental parameters include light intensity and water turbidity, and the value range is generally between 0.5 and 1.5 to reflect the changes in the impact of the environment on light propagation; κ is the adjustment frequency factor, and the value range is usually set to 0.1 to 1.0, which is used to control the response speed of the spectral matching function to environmental changes; Γ(t) is a composite function of light intensity and turbidity, which comprehensively evaluates the changes in ambient light and water transparency, and the value range is set to 100 to 1000 to cover a wide range of scenarios from low light and low transparency to high light and high transparency.

[0100] The current environmental parameters are as follows: the light intensity is 800, and the turbidity index is 0.8. In this case, set γ(t) = 1.2, κ = 0.5, and Γ(t) = 800 + (0.8 × 100) = 880. Substitute these parameters into the formula for calculation:

[0101]

[0102] Set the time interval T = 10 seconds and consider (Because the change of environmental parameters in a short time can be approximated as a constant), then:

[0103] Φ(t)=∫0 10 [1.2·cos(0.5·t)]dt

[0104] The calculation results are:

[0105]

[0106] The adjusted spectral matching function Φ(t)≈2.3 will be used to update the parameters in the spectral matching algorithm, especially the weight coefficient and attenuation characteristics in the depth calculation formula, so as to adjust the fusion effect of multi-band data in real time. Through this dynamic adjustment process, the system can still maintain efficient correction of distorted areas and accurate depth reconstruction in complex and changing underwater environments, ensuring the high consistency of the reconstructed 3D image in both overall and detailed aspects.

[0107] Finally, in order to further improve the accuracy of the image, the nonlinear fusion model is used to integrate the spectral information of different bands into a multi-dimensional feature space, and multi-level feature mapping and deep information extraction are performed through a deep learning network. Visible light, infrared and ultraviolet spectral data collected by multispectral imaging technology are integrated into a high-dimensional feature space. In this process, the formula for nonlinear fusion and deep extraction is used:

[0108]

[0109] Among them, Z(x,y) represents the depth information after nonlinear fusion at the pixel point (x,y), Λ is the spectral range, δ j is the weight coefficient of different bands, reflecting the contribution of each band in the extraction of depth information, and the value range is set to 0.1 to 1.0; ω j is the frequency factor, which controls the frequency characteristics of the spectral response and has a value range of 0.05 to 0.5; and T j (λ) is the nonlinear spectral response function of the jth band, which describes the spectral response characteristics of each band.

[0110] m = 4 spectral bands are used, corresponding to the visible light, near infrared and ultraviolet ranges, and the wavelength range Λ is set to 300nm to 900nm to cover the spectral data from ultraviolet to near infrared. The specific weight coefficients δ1 = 0.8, δ2 = 0.7, δ3 = 0.6, δ4 = 0.5 are set. These values ​​are determined according to the underwater penetration and image contribution of each band, reflecting the influence of different bands on depth information. Frequency factor ω j The values ​​are set as ω1 = 0.1, ω2 = 0.2, ω3 = 0.3, and ω4 = 0.4. These values ​​are used to adjust the response frequency of each band during the fusion process to adapt to the performance of different spectral characteristics in the underwater environment.

[0111] Substitute these parameters into the formula to calculate the depth information for a certain pixel (x=50, y=100):

[0112]

[0113] The spectral response function of each band is set to T1(300nm)=0.9, T2(500nm)=0.85, T3(700nm)=0.8, T4(900nm)=0.75 at wavelengths of 300nm, 500nm, 700nm, and 900nm, respectively. These data are put into the formula and calculated item by item:

[0114] Z(50,100)=(0.8·0.9·sin(0.1·300))+(0.7·0.85·sin(0.2·500))+(0.6·0.8·sin(0.3·700))+(0.5·0.75·sin(0.4·900))

[0115] =(0.8·0.9·sin(30))+(0.7·0.85·sin(100))+(0.6·0.8·sin(210))+(0.5·0.75·sin(360))

[0116] ≈(0.72·0.5)+(0.595·0.984)+(0.48·-0.5)+(0.375·0)

[0117] =0.36+0.58548-0.24+0=0.70548

[0118] Through this calculation process, the nonlinear fusion depth information of the pixel point (50,100) is 0.70548 meters. This nonlinear fusion and depth extraction method significantly improves the depth accuracy of underwater 3D reconstruction by multi-dimensionally processing spectral data of different bands.

[0119] Embodiment 2:

[0120] See attached Figure 3 In this embodiment, in order to improve the accuracy of depth measurement during the three-dimensional reconstruction of underwater structures, depth compensation technology is used. By establishing an underwater optical transmission model, the scattering path and absorption coefficient of light during underwater propagation are analyzed, and the nonlinear effect of scattering and absorption on depth measurement is dynamically calculated. This process uses the path integral formula to simulate the propagation characteristics of light, and the formula is:

[0121] I(x,y,z)=∫0 L [η(λ)·e -α(λ)·s ·cos(θ(s))]ds

[0122] In this formula, I(x,y,z) represents the light intensity at the point (x,y,z), which reflects the energy transmission of light in an underwater environment; λ is the wavelength of light, and α(λ) is the absorption coefficient, which represents the attenuation degree of light at a specific wavelength. It usually ranges from 0.01 to 0.1 to reflect the attenuation characteristics of light of different wavelengths in water. For example, the absorption coefficient of blue light can be set to 0.02, while the absorption coefficient of red light can be set to 0.08 due to its rapid attenuation in water; the spectral response function η(λ) of the scattering path describes the performance of light of different wavelengths after scattering in water, and its value range is usually set to 0.5 to 1.5 to reflect the different scattering degrees of light underwater. The response of blue light is set to 1.2 and that of red light to 0.7; θ(s) represents the angular change of the scattering path, and its value can be set between 0 and 90 degrees according to the flow of the water body and the environment. The integral path L represents the length of the propagation path of light in the water body, and its value usually ranges from 1 to 10 meters, depending on the depth and clarity of the underwater environment.

[0123] Assume that a certain pixel point (x=100, y=150, z=3) is located on the surface of the target underwater structure, the wavelength λ of the light is 500nm (blue light), the absorption coefficient α(500nm)=0.02, the scattering response η(500nm)=1.2, and the propagation path of the light is L=5 meters, and the scattering angle change θ(s)=30 degrees. Substitute these parameters into the formula for calculation:

[0124] I(100,150,3)=∫0 5 [1.2·e -0.02·s ·cos(30°)]ds

[0125] Simplified calculation:

[0126]

[0127] The light intensity at this point is about 4.946, which reflects the attenuation of light after propagation and scattering 5 meters underwater. This light intensity value will be used for subsequent calculations of depth compensation, and the measurement accuracy can be further optimized by adjusting the absorption coefficient and response function.

[0128] In order to cope with the influence of different depths and water turbidity on light propagation, the parameters of the correction model are dynamically adjusted based on the real-time monitoring of underwater environmental parameters, and a multi-level correction mechanism is used to optimize depth measurement. In shallow areas, due to the low turbidity of the water body, the scattering and absorption of light are small, and the linear correction of the light scattering coefficient is used. In deep or high turbidity areas, the nonlinear scattering and absorption of light are more significant, so it is necessary to use depth compensation based on the nonlinear scattering path to correct the depth measurement error. This process is achieved through the following formula:

[0129]

[0130] In this formula, D(z) represents the corrected depth value at depth z, which is dynamically calculated through a multi-level correction mechanism. The part represents the shallow linear correction, where γ n is the correction factor, which reflects the correction strength and is usually in the range of 0.5 to 1.5; β n It is a depth-dependent attenuation factor used to describe the degree of shallow light attenuation, with a value range of 0.01 to 0.1. It is used to process nonlinear depth compensation in complex water environments, where δ n (z) represents the nonlinear correction function, which reflects the change of light scattering path in deep layers or high turbidity conditions. It is usually set as a function that is sensitive to changes in depth and can capture nonlinear characteristics.

[0131] Assuming that the current water environment monitoring shows that the light scattering coefficient in the shallow area (within 1 meter) is low, set γ1 = 1.0, β1 = 0.02; while in the deep layer (5 meters) due to the increase in water turbidity, the nonlinear effect is significant, set γ2 = 1.2, β2 = 0.05, and the nonlinear correction term δ2(z) = 0.1z 2 It is necessary to calculate the corrected depth value D(5) at the depth z=5 meters.

[0132] Substituting these parameters into the formula:

[0133]

[0134] Expand calculation:

[0135] For shallow correction (n=1):

[0136]

[0137] For deep correction (n=2), nonlinear compensation terms are included:

[0138]

[0139] Add the results of the two parts:

[0140] D(5)=4.76+5.3088=10.0688

[0141] Therefore, the corrected depth value at a depth of 5 meters is 10.0688 meters. This dynamic calculation through a multi-level correction mechanism can accurately compensate for the nonlinear effects of different depths and water turbidity on light propagation.

[0142] In order to ensure the real-time and accuracy of depth compensation and reconstruction, a parallel processing architecture is used to perform scatter compensation and depth reconstruction in real time. Through the depth error feedback mechanism, the system can dynamically adjust the compensation parameters to ensure a high degree of coordination between the compensation and reconstruction processes. This synchronous compensation and reconstruction process is achieved through the following formula:

[0143]

[0144] Among them, E(t) represents the depth error after real-time compensation, reflecting the error correction effect of the entire compensation and reconstruction process. ξ(t) in the formula is the scattering compensation coefficient, which is used to adjust the compensation intensity in real time. The value range is generally 0.5 to 1.5 to meet the scattering compensation requirements in different environments; ω is the synchronization frequency factor, which determines the coordination frequency between compensation and reconstruction. The value range is set to 0.1 to 1.0 Hz. is the reconstruction feedback adjustment term, ζ(t) reflects the error feedback in the reconstruction process, and the compensation effect is further optimized by performing differential operation on the real-time feedback in the reconstruction process.

[0145] Assume that the current environmental monitoring shows that the underwater illumination and turbidity have certain fluctuations, resulting in the need for real-time adjustment of scatter compensation. Set the scatter compensation coefficient ξ(t) = 1.2, the frequency factor ω = 0.5, and the reconstruction feedback function ζ(t) = 0.1t 2 , where the form of ζ(t) is chosen to capture the gradually accumulated error in the compensation process and dynamically adjust the compensation intensity through its derivative.

[0146] Substituting these parameters into the formula, calculate the depth error compensation effect at time T = 10 seconds:

[0147]

[0148] Expand calculation:

[0149]

[0150] Therefore, the depth error after real-time compensation within 10 seconds is 13.08088 meters. This real-time compensation and reconstruction synchronization mechanism can continuously optimize the compensation parameters through parallel processing architecture and error feedback adjustment to ensure the accuracy of depth reconstruction when the environment changes.

[0151] Embodiment 3:

[0152] In this embodiment, when continuing the three-dimensional reconstruction of underwater structures, in order to accurately capture the changes in the underwater environment and dynamically adjust the three-dimensional reconstruction parameters, a solution for generating an environmental characteristic model is used. First, in the underwater environment, sensors are used to monitor multidimensional parameters such as light, turbidity, and temperature in real time. Using dynamic environmental feature extraction technology, these multi-source data are nonlinearly analyzed and fused to form a preliminary environmental characteristic model. In order to adapt to the dynamic changes in the environment during the three-dimensional reconstruction process, the system uses an adaptive algorithm based on real-time data to adjust the model, and dynamically calculates the response parameters of the environmental characteristic model through the following formula:

[0153]

[0154] Among them, M(t) represents the dynamic response parameter of the environmental characteristic model at time t, reflecting the impact of the environment on the 3D reconstruction process. The parameter α(t) represents the weight coefficient of different environmental parameters including illumination, turbidity, and temperature. These weights are dynamically adjusted according to real-time monitoring data, usually ranging from 0.5 to 1.5 to reflect the importance of each environmental parameter under different conditions; ω is the frequency factor of the model response, usually set to 0.1 to 1.0 Hz, representing the periodic frequency of environmental changes; β(t) is the characteristic function that describes the nonlinear changes of environmental parameters, and its value can dynamically reflect the changing trend of the underwater environment.

[0155] In a specific area of ​​the research site, the sensor measures the current light intensity to be 800 Lux, the turbidity to be 0.7 NTU, and the water temperature to be 15 degrees Celsius. Based on these data, the light weight coefficient α is set. 光照 =1.2, turbidity weight coefficient α 浑浊度 =1.0, the temperature weight coefficient α 温度 =0.8. Set ω=0.5Hz, β(t)=0.05t 2 , where the form of β(t) is used to capture the nonlinear characteristics of environmental changes, such as the fluctuation of illumination over time and the gradual change of turbidity.

[0156] Substituting these data into the formula to calculate the environmental characteristic model response at time T = 10 seconds:

[0157]

[0158] Expand calculation:

[0159]

[0160] Therefore, the environmental feature model response within 10 seconds is 7.5674. This result reflects the dynamic regulation of the combined effects of light, turbidity and temperature on the 3D reconstruction process under current environmental conditions.

[0161] To further improve the construction of the environmental characteristic model, we need to model illumination, turbidity and temperature locally to capture the characteristics of different depths and regions, and integrate these local models into a global environmental characteristic model. To this end, we use a multi-level optimization algorithm for dynamic adjustment to ensure that the environmental characteristic model can reflect the current underwater environmental changes in real time. This optimization process is achieved through the following formula:

[0162]

[0163] Among them, G(x,y,z) is the global environment feature model at the spatial position (x,y,z), γ n It is a local environmental characteristic parameter, which indicates the initial value of environmental characteristics in a specific area or depth. It usually ranges from 0.5 to 1.5 to reflect the influence of local characteristics under different environmental conditions. n is the depth-related environmental impact factor, which describes the changing trend of environmental parameters with depth. Its value range is 0.01 to 0.1 to reflect the attenuation or enhancement of environmental parameters at different depths; ∈ n (z) is a nonlinear optimization term, which is used to further correct the accuracy of the model and describes the nonlinear change characteristics of environmental parameters.

[0164] At the underwater structure research site, the illumination, turbidity and temperature at different depths and areas were monitored and modeled. At a depth of Z = 3 meters, the measured illumination intensity, turbidity and temperature were 700 Lux, 0.9 NTU and 14 degrees Celsius, respectively. Based on these data, the local characteristic parameter γ of illumination was set 光照 =1.1, local characteristic parameter γ of turbidity 浑浊度 = 0.9, local characteristic parameter γ of temperature 温度 =1.3. The depth-related influencing factors are set as δ 光照 =0.02, δ 浑浊度 =0.05 and δ 温度 =0.03, nonlinear optimization term∈ n (z) = 0.1z 2 , which is used to capture the nonlinear change trend of various environmental parameters.

[0165] Substituting these data into the formula, we can calculate the response of the global environmental characteristic model at point (x=100, y=150, z=3):

[0166]

[0167] Expand the calculation, for lighting:

[0168]

[0169] For turbidity:

[0170]

[0171] For temperature:

[0172]

[0173] Add the partial results together:

[0174] G(100,150,3)=3.219+2.518+4.631=10.368

[0175] The global environmental characteristic model response at a depth of 3 meters is 10.368. This shows that by modeling and integrating local features into a global model and applying a multi-level optimization algorithm, the impact of environmental features at different depths and regions can be dynamically captured and adjusted.

[0176] In order to improve the system's ability to respond to environmental changes, an intelligent sensor data fusion mechanism is used to deeply fuse the light, turbidity and temperature data from different sensors, and predict future environmental changes through time series analysis. The core of this step is to provide predictive adjustments for the reconstruction process through real-time data fusion and environmental change prediction models. The core formula for data fusion and prediction is:

[0177]

[0178] Among them, P(t) represents the environmental change prediction parameter at time t, reflecting the future environmental trend. λ(t) is the fusion weight of different sensor data, which determines the influence of each sensor in the prediction result. The value range is 0.5 to 1.5 to flexibly adjust the importance of each sensor. κ is the frequency factor of the time series, which is usually set to 0.1 to 1.0 Hz, reflecting the periodicity and trend of environmental changes. θ(t) is the prediction correction term, which is used to correct the prediction results by analyzing the historical trend of sensor data to ensure accurate prediction of future environmental changes.

[0179] The data measured by the sensor used in the setting are: the current light intensity is 800 Lux, the turbidity is 0.8 NTU, and the water temperature is 15 degrees Celsius. Set the fusion weight of the light to λ 光照 =1.2, the fusion weight of turbidity is λ 浑浊度 =1.0, the fusion weight of temperature is λ 温度 =0.9. The frequency factor of the time series κ = 0.5, and the forecast correction term θ(t) = 0.05t 2 , used to capture nonlinear trends such as increasing turbidity and light fluctuations.

[0180] Substituting these data into the formula, calculate the predicted parameters of environmental change at time T = 10 seconds:

[0181]

[0182] Expand calculation:

[0183]

[0184] Therefore, the predicted parameter of environmental changes within 10 seconds is 3.0822. This result shows that the system can dynamically adjust the key parameters in the 3D reconstruction process through the fusion of intelligent sensor data and accurate prediction of environmental changes, ensuring that the reconstruction remains accurate and stable in the future environmental changes.

[0185] Embodiment 4:

[0186] In this embodiment, in the underwater three-dimensional reconstruction process of underwater structure research, accurate identification of local distortion areas is crucial to improving reconstruction quality. The local distortion area identification method used constructs a high-dimensional feature space by integrating environmental parameters such as illumination, turbidity, temperature, and image brightness, texture, and edge features, thereby identifying local distortion areas caused by environmental changes. This method generates a multi-dimensional matching matrix by analyzing the nonlinear relationship between environmental parameters and image features to achieve accurate distortion identification. The specific mathematical representation is as follows:

[0187]

[0188] Among them, F(x,y,t) represents the matching function value at the spatial position (x,y) and time t, representing the comprehensive influence between environmental parameters and image features. Parameter α i (t) is the weight of environmental parameters such as illumination, turbidity, and temperature, which controls the contribution of each environmental factor to image distortion. Its value range is 0.5 to 1.5 to flexibly adjust the influence of environmental factors; g i (x, y, t) is the nonlinear response function of the image feature, which is used to capture the impact of environmental changes on image quality and reflect the changes in image brightness, texture, and edges; the correction term β i (t) is used to adjust the model to improve the accuracy of distortion identification and capture the temporal variation trend.

[0189] In the underwater environment at the research site, the current light intensity measured by the sensor is 750 Lux, the turbidity is 1.0 NTU, and the temperature is 16 degrees Celsius. Based on these data, the light weight α is set 光照 =1.1, turbidity weight α 浑浊度 =1.2, the temperature weight α 温度 =0.9. Image feature response function gi (x, y, t) is set as a nonlinear function according to the changes in brightness, texture and edge, for example, g1(x, y, t) = 0.8x, g2(x, y, t) = 0.7y 2 and g3(x,y,t)=0.6t. Correction term β i (t)=0.05t 2 , which is used to further improve the accuracy of the model at different time points.

[0190] Substitute these data to calculate the matching function value at time T = 10 seconds and spatial position (x = 50, y = 75):

[0191]

[0192] Expand calculation:

[0193]

[0194]

[0195] Therefore, within 10 seconds, the matching function value at the spatial position (50,75) is 47749. This result accurately reflects the combined effects of illumination, turbidity, and temperature on image distortion by comprehensively matching environmental parameters with image features, thereby identifying local distorted areas.

[0196] As the identification of local distortion areas is gradually carried out, the accuracy and adaptability of distortion identification are further improved through the intelligent distortion detection model based on the deep learning algorithm. The model learns different types of distortion features by training environmental data and a large number of image samples, thereby establishing a multi-level distortion identification system. The adaptive learning process is expressed by the following formula:

[0197]

[0198] Among them, S(t) represents the distortion detection result at time t, reflecting the comprehensive judgment of the current environment and image features. Parameter γ j (t) is the weight factor of the distortion feature, which measures the importance of each type of distortion feature in detection. The value range is generally 0.5 to 1.5 to adjust the impact of different distortion features on the model; h j (t) is the corresponding characteristic response function, which is used to describe the dynamic behavior of the distortion feature and reflects the impact of environmental changes on image distortion; the correction term δ j (t) is used to adjust the sensitivity and adaptability of the model to further improve the accuracy of distortion identification.

[0199] The underwater environment is reconstructed in three dimensions at the underwater structure research site. The intelligent distortion detection model learns and adjusts through real-time data monitored by sensors and historical image samples. The current measured environmental parameters are set as follows: light intensity 800 Lux, turbidity 1.2 NTU, and water temperature 17 degrees Celsius. Set the distortion weight of light γ 光照 =1.0, distortion weight of turbidity γ 浑浊度 =1.3, temperature distortion weight γ 温度 =0.8. The characteristic response function is set as: h1(t) = 0.9sin(0.2t) to describe the illumination distortion characteristics, h2(t) = 1.1cos(0.3t) to describe the turbidity distortion characteristics, and h3(t) = 0.7sin(0.1t) to describe the temperature-related distortion characteristics. Correction term δ j (t) = 0.1t 2 Used to adjust the model to reflect changes in nonlinear distortion characteristics in the time series.

[0200] Substitute these data to calculate the distortion detection results for time T = 10 seconds:

[0201]

[0202] Expand calculation:

[0203]

[0204]

[0205] Therefore, the distortion detection result within 10 seconds is 15.5713. Through this adaptive deep learning detection model, the system can adjust the comprehensive judgment of the environment and image features in real time, accurately identify various types of distortion features, and improve the sensitivity and accuracy of distortion recognition through continuous learning and optimization.

[0206] In order to further improve the synchronous analysis and processing of environmental parameters and image features, a collaborative working mechanism of multi-dimensional matching and intelligent detection model is adopted. This collaborative detection mechanism combines the joint evaluation of the matching matrix and the intelligent model to perform in-depth analysis and positioning of local distortion areas, achieving high-precision distortion identification and correction. The collaborative detection process is carried out through the following formula:

[0207]

[0208] Among them, C(t) represents the collaborative detection result at time t, which is a comprehensive evaluation of the coupling relationship between environmental parameters and image features. The parameter λ(t) is the weight of collaborative detection, which measures the interaction between environmental and image features in distortion detection. It usually ranges from 0.5 to 1.5 to adjust the impact of different factors on the detection results; ω is the frequency factor of the time series, ranging from 0.1 to 1.0 Hz, which is used to capture the periodic changes of environmental and image features; the correction term θ(t) is used to dynamically adjust the detection results, reflecting the system's ability to adapt to real-time environmental changes.

[0209] Continue to use sensors to monitor the environmental data near the underwater structure. The current measured light intensity is 850 Lux, turbidity is 1.1 NTU, and water temperature is 16.5 degrees Celsius. Based on these data, set the collaborative detection weight λ of light 光照 =1.2, turbidity collaborative detection weight λ 浑浊度 =1.0, the collaborative detection weight of temperature λ 温度 =0.9. Frequency factor ω = 0.4 Hz, reflecting the periodic changes of the environment and image characteristics. Correction term θ(t) = 0.07t 2 Used to dynamically adjust the test results to accurately reflect the changing trends over time.

[0210] Substituting these parameters into the formula, calculate the collaborative detection result at time T = 10 seconds:

[0211]

[0212] Expand calculation:

[0213] C(10)=∫0 10 [1.0·cos(0.4t)+0.14t]dt

[0214]

[0215] Therefore, the collaborative detection result within 10 seconds is 5.108. This shows that the system can effectively integrate the changes in environmental parameters and image features through the collaborative work of multi-dimensional matching and intelligent detection models, and adjust the distortion detection strategy in real time, so as to accurately identify and locate the distorted area.

[0216] Embodiment 5:

[0217] In this embodiment, in the process of underwater 3D reconstruction of underwater structure research, in order to ensure the high quality and accuracy of the image, a scheme of gradually correcting the distorted area is adopted. This scheme combines the dynamic feedback of environmental parameters and image features to perform multiple rounds of iterative corrections on the distorted area to continuously optimize the reconstruction effect. The mathematical expression of the optimization process is:

[0218]

[0219] Among them, L(t) represents the nonlinear loss function at time t, reflecting the error size in the current correction process. Parameter α i is the weight factor of environmental parameters and image features, which is used to measure the impact of each factor on distortion. Its value range is 0.5 to 1.5 to dynamically adjust the weight of different factors. i (x, y, t) is the nonlinear response function of a specific distortion area, which describes the impact of environmental changes on the image, such as the different responses of illumination and turbidity to image distortion; β(t) is the correction term used for optimization, which realizes dynamic adjustment of the error through differential operation and improves the accuracy of correction.

[0220] In the reconstruction process of underwater structures, the currently monitored environmental parameters are light intensity 820 Lux, turbidity 1.3 NTU, and water temperature 18 degrees Celsius. Based on these environmental data, the light weight α is set. 光照 =1.0, turbidity weight α 浑浊度 =1.2, the temperature weight α 温度 =0.8. For the nonlinear response function of a specific distortion area, Φ1(x,y,t)=0.9sin(0.3t), Φ2(x,y,t)=1.1cos(0.2t), Φ3(x,y,t)=0.7sin(0.1t), which are used to describe the effects of illumination, turbidity and temperature on the image, respectively. Correction term β(t)=0.1t 2 , which is used to optimize the distortion correction process.

[0221] Substituting these parameters into the formula, we can calculate the nonlinear loss function at time T = 10 seconds:

[0222]

[0223] Expand calculation:

[0224]

[0225] Therefore, the nonlinear loss function within 10 seconds is 21.9954. This result shows that the system can effectively reduce image distortion and improve the accuracy of 3D reconstruction by gradually correcting the distorted area and combining dynamic feedback of environmental parameters and image features.

[0226] In order to optimize the reconstruction quality and accuracy of the image, a hierarchical image processing scheme is adopted. This scheme divides the image into multiple scale levels and performs optimization processing at different scales to balance the global structure and local details to ensure high-precision reconstruction in complex environments. The mathematical expression of the multi-scale optimization process is:

[0227]

[0228] Among them, G(x,y,z) represents the global optimization function at the spatial position (x,y,z), reflecting the optimization effect at different levels. j is the optimization parameter at different scales, usually ranging from 0.5 to 1.5, to adjust the optimization intensity of each scale; δ j is a scale-related attenuation factor, describing how the optimization effects at different levels change with depth. Its value range is 0.01 to 0.1, indicating the attenuation of optimization from global to local; ∈ j (z) is the optimization correction term, which is used to further balance the global structure and local details, and realize dynamic adjustment of each level through differential operation.

[0229] In the 3D reconstruction of underwater structures, the image is processed at multiple scales and the depth level is set to Z = 5 meters. The illumination, turbidity and temperature at different scales are optimized. Based on these parameters, the optimization parameter γ for illumination is set. 光照 =1.1, the optimized parameter of turbidity γ 浑浊度 =1.3, the optimized parameter of temperature γ 温度 =0.9. For the depth-dependent attenuation factor, set δ 光照 =0.02, δ 浑浊度 =0.05, δ 温度 =0.03. Correction term ∈ j (z) = 0.1z 2 , which is used to dynamically balance the optimization effects at each level.

[0230] Substituting these data into the formula, the global optimization function is calculated at a depth of Z = 5 meters:

[0231]

[0232] Expand calculation:

[0233] For lighting optimization:

[0234]

[0235] For turbidity optimization:

[0236]

[0237] For temperature optimization:

[0238]

[0239] Add the partial results together:

[0240] G(100,150,5)=5.264+5.765+6.679=17.708

[0241] Therefore, the global optimization function at a depth of 5 meters is 17.708. This result shows that through layered processing and combining optimization at different scales, the system can effectively balance global structure and local details, ensuring high accuracy and quality in the 3D reconstruction process.

[0242] Embodiment 6:

[0243] In this embodiment, in the underwater 3D reconstruction process of underwater structure research, in order to accurately extract high-level features in the image, especially nonlinear distortion features caused by environmental changes, a convolutional neural network (CNN) is introduced to gradually extract and analyze complex features in the image through multi-layer convolution and pooling operations. This process is performed through the following mathematical expression for feature extraction:

[0244]

[0245] In this formula, F(x,y,t) represents the feature mapping value at the spatial position (x,y) and time t, which reflects the feature changes of the image under different environmental conditions. i is the weight coefficient of different convolutional layers, ranging from 0.5 to 1.5, which is used to adjust the influence of each convolutional layer on feature extraction; σ is the activation function, usually ReLU or Sigmoid, to capture complex feature relationships through nonlinear mapping; W i is the weight of the convolution kernel, · represents the weighted sum of the convolution operation on the input feature g(x,y,t), and b i It is a bias term, usually ranging from -0.5 to 0.5, and is used to adjust and correct the distribution of features.

[0246] CNN is used for feature extraction in underwater images of underwater structure research sites. The current environmental data is set to light intensity 800Lux, turbidity 1.2NTU, water temperature 17 degrees Celsius, and the input function of image features is set to g(x,y,t)=0.8x+0.6y-0.2t. The number of convolution layers is set to N=3, and the convolution kernel weights are W1=0.9, W2=1.1, W3=0.8, and the corresponding weight coefficients are α1=1.2, α2=0.8, α3=1.0. The bias terms are set to b1=-0.1, b2=0.2, b3=0.0, and the activation function is selected as ReLU, that is, σ(x)=max(0,x).

[0247] Substituting these parameters into the formula, the feature map value at time T = 10 seconds and spatial position (x = 50, y = 75) is calculated:

[0248]

[0249] Expand calculation:

[0250] For convolutional layer 1:

[0251]

[0252] After activation:

[0253]

[0254] For convolutional layer 2:

[0255]

[0256] For convolutional layer 3:

[0257]

[0258]

[0259] Add the results of each layer:

[0260] F(50,75,10)=906+740.8+672=2318.8

[0261] Therefore, at time 10 seconds, the feature map value of spatial position 50,75 is 2318.8. This result shows that through the multi-layer convolution and pooling operations of CNN, the system is able to gradually extract high-level features in the image, especially those nonlinear distortion features caused by environmental changes.

[0262] To further optimize the quality and authenticity of the image, the Generative Adversarial Network (GAN) architecture is introduced. GAN consists of two parts: the generator and the discriminator. The generator uses the features extracted by the convolutional neural network (CNN) to generate high-quality images, while the discriminator evaluates the authenticity of the generated images. The adversarial loss function of GAN is used to optimize the performance of the generator and the discriminator, and its expression is as follows:

[0263]

[0264] Among them, L(G,D) represents the adversarial loss function between the generator G and the discriminator D, which aims to optimize the performance of the two networks in adversarial training. Represents the expectation of the real image x, and the discriminator D is optimized by judging whether the input image comes from real data; It is to evaluate the images output by the generator. The generator G deceives the discriminator by generating realistic images, thereby continuously improving its generation ability.

[0265] In practice, the generator learns to generate high-quality images from features previously extracted by CNN. The current feature map value is set to F(50,75,10)=2318.8, and the generator is trained with this as input. The setting of the convolution layer includes the initial noise distribution z sampling as a normal distribution with mean 0 and standard deviation 1. The initial parameters of the discriminator are set to weights between 0.8 and 1.2, and the learning rate is 0.001.

[0266] Substituting these parameters for adversarial training, first, the discriminator evaluates the probability of the real image as:

[0267]

[0268] Then, the generator generates an image based on the input noise z and is evaluated by the discriminator:

[0269]

[0270] Add the two expected values ​​to get the value of the adversarial loss function:

[0271] L(G,D)=-0.105+(-0.511)=-0.616

[0272] During adversarial training, the generator continuously adjusts weights and biases to optimize the quality of generated images and make them more similar to real images; the discriminator continuously distinguishes between real and generated images to optimize its own judgment ability. After multiple rounds of iterations, the adversarial relationship between the generator and the discriminator gradually reaches a balance, and the images output by the generator become realistic.

[0273] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for three-dimensional reconstruction of an underwater environment, characterized in that The following steps are involved: S1, projecting the optimized spectral pattern onto the target surface, capturing the reflected light with a camera, and inferring the three-dimensional structure of the object surface through the deformation of light; in: S1.1, projecting multi-band structured light patterns in an underwater environment while capturing images using a multispectral camera; S1.2, through the spectrum matching algorithm, the images of different bands are fused, the depth information is extracted and the basic data for 3D reconstruction is generated; S1.

3. Use depth compensation technology to correct the depth error caused by underwater scattering and absorption; S2, dynamically adjust the image correction strategy by real-time monitoring of underwater environmental parameters; in: S2.

1. Use underwater sensors to monitor environmental parameters including light, turbidity, and temperature in real time and generate environmental characteristic models; S2.2, adaptively adjust image acquisition and processing parameters including exposure, contrast and color balance according to the environmental feature model; S2.3, through intelligent correction algorithm, detect the distorted areas in the image caused by environmental changes, and perform local correction and overall optimization; S3, by constructing a multi-level image pyramid model, first capture the overall structure at a large scale, and then gradually enhance the texture details at a small scale; in: S3.

1. Generate a multi-scale image pyramid to decompose the structural information of the three-dimensional image layer by layer from coarse to fine; S3.2, using a generative adversarial network to enhance details and restore textures at each layer; S3.

3. Through adaptive detail enhancement technology, the key areas are refined to generate underwater three-dimensional images.

2. The method for three-dimensional reconstruction of an underwater environment according to claim 1, characterized in that The spectrum matching algorithm construction process includes: S1. Using multispectral imaging technology, multiple spectral information of visible light, infrared light and ultraviolet light is integrated into a multidimensional spectral feature matrix. The depth information of each pixel is calculated through the spectral response difference of each band. The calculation formula is expressed as: Where D(x,y) represents the depth value at the pixel point (x,y), λ is the spectral wavelength, S i (λ) is the spectral response of the i-th band, α i and β i (λ) is the coefficient and absorption characteristics used to weigh the contributions of different bands; S2. Adjust the parameters of the spectral matching algorithm according to the real-time changes of the underwater environment; dynamically update the matching function to ensure the accuracy of the image by monitoring the illumination and turbidity environmental parameters; the correction process is achieved through the adaptive adjustment function: Among them, Φ(t) is the spectral matching function after dynamic adjustment, γ(t) represents the influence of environmental parameters on the correction function, κ is the adjustment frequency factor, and Γ(t) is the composite function of ambient light intensity and turbidity; S3. Utilize nonlinear fusion models to integrate spectral information of different bands into a multi-dimensional feature space, and perform multi-level feature mapping and deep information extraction through a deep learning network.

3. The method for three-dimensional reconstruction of an underwater environment according to claim 1, characterized in that The depth compensation technology adopts the following scheme: S1. By establishing an underwater optical transmission model, the scattering path and absorption coefficient of light during underwater propagation are analyzed, and the nonlinear effects of scattering and absorption on depth measurement are dynamically calculated; simulation is performed using the path integral formula, which is expressed as follows: I(x,y,z)=∫0 L [η(λ)·e -α(λ)·s ·cos(θ(s))]ds Among them, I(x,y,z) represents the light intensity at the point (x,y,z), which describes the energy of light propagating in an underwater environment; λ is the wavelength of light, and α(λ) is the absorption coefficient, which represents the degree of attenuation of light at a specific wavelength; S2. Dynamically adjust the parameters of the correction model according to the underwater environmental parameters monitored in real time; through a multi-level correction mechanism, linear correction of the light scattering coefficient is used in shallow areas, and depth compensation based on nonlinear scattering paths is used in deep and high turbidity areas; S3. Through the parallel processing architecture, scatter compensation and depth reconstruction are performed in real time, and the compensation parameters are dynamically adjusted using the depth error feedback mechanism.

4. The method for three-dimensional reconstruction of an underwater environment according to claim 1, characterized in that The process of generating the environmental feature model comprises: S1. First, in the underwater environment, real-time monitoring of multi-dimensional parameters such as light, turbidity, and temperature is performed, and dynamic environmental feature extraction technology is used to perform nonlinear analysis and fusion of multi-source data; S2, then locally model the illumination, turbidity and temperature to capture the characteristics of different depths and regions; integrate the local models into a global environmental feature model and dynamically adjust it through a multi-level optimization algorithm; S3. Finally, the data from different sensors are deeply integrated through the intelligent sensor data fusion mechanism, and future environmental changes are predicted through time series analysis. The core formula for data fusion and prediction is: Among them, P(t) represents the environmental change prediction parameter at time t, λ(t) is the fusion weight of different sensor data, which determines the influence of each sensor in the prediction result, κ is the frequency factor of the time series, which reflects the periodicity and trend of environmental changes, and θ(t) is the prediction correction term, which is used to correct the prediction result.

5. The method for three-dimensional reconstruction of an underwater environment according to claim 1, characterized in that The construction method steps adopted by the intelligent correction algorithm include: S1. Identify local distortion areas caused by environmental changes through multi-dimensional matching of environmental parameters and image features; and distinguish and locate distortion areas based on an intelligent distortion detection model; S2, use nonlinear optimization method to gradually correct the distorted area; at the same time, combine multi-scale optimization technology to perform layered processing on the entire image; S3. Use the convolutional neural network (CNN) to extract and analyze the features of the distorted area, and then reconstruct the image through the generative adversarial network (GAN).

6. The method for three-dimensional reconstruction of an underwater environment according to claim 5, characterized in that The construction process of the local distortion area identification method includes: S1, by comprehensively matching the environmental parameters of illumination, turbidity, and temperature with the image features of brightness, texture, and edge, a high-dimensional feature space is formed to identify the local distortion areas caused by environmental changes; S2. Based on the deep learning algorithm, the intelligent distortion detection model learns different types of distortion features through training of environmental data and image samples, and establishes a multi-level distortion recognition system; S3. Finally, through the collaborative work of multi-dimensional matching and intelligent detection model, the environmental parameters and image features are analyzed and processed synchronously; and the collaborative detection mechanism combines the joint evaluation of the matching matrix and the intelligent model to perform in-depth analysis and positioning of the local distortion area; and the collaborative detection formula is as follows: Among them, C(t) represents the collaborative detection result at time t, which is a comprehensive evaluation of the coupling relationship between environmental parameters and image features; the parameter λ(t) is the weight of collaborative detection, which measures the interaction between environmental and image features in distortion detection; ω is the frequency factor of the time series, which is used to capture the periodic changes of environmental and image features; the correction term θt is used to dynamically adjust the detection results.

7. The method for three-dimensional reconstruction of an underwater environment according to claim 5, characterized in that The stepwise correction of the distorted area is to perform multiple rounds of iterative correction on the distorted area by combining dynamic feedback of environmental parameters and image features; The hierarchical processing of the image is to divide the image into multiple scale levels and perform optimization processing at different scales; the expression of the multi-scale optimization process is as follows: Among them, G(x,y,z) represents the global optimization function at the spatial position (x,y,z), γ j is the optimization parameter at different scales, δ j is a scale-dependent attenuation factor, describing the optimization effects at different levels, ∈ j (z) is the optimization correction term, which is used to balance the global structure and local details.

8. The method for three-dimensional reconstruction of an underwater environment according to claim 5, characterized in that The feature extraction and analysis described above introduces a convolutional neural network (CNN) to gradually extract high-level features in the image through multi-layer convolution and pooling operations, including nonlinear distortion features caused by environmental changes; The generator of the generative adversarial network GAN generates high-quality images using the features extracted by CNN, while the discriminator evaluates the authenticity of the generated images; the expression of the loss function is as follows: Among them, L(G,D) represents the adversarial loss function between the generator G and the discriminator D; is the expectation of the real image x, and the discriminator D is optimized by judging whether the input image comes from real data or the generator; It is to evaluate the images output by the generator, which deceives the discriminator by generating images.

Images