Ecological restoration project management method and system based on three-dimensional live-action model

Abstract

The invention discloses an ecological restoration project management method and system based on a three-dimensional live-action model, and relates to the technical field of ecological restoration. The method comprises the following steps: firstly, acquiring image data of a mine area through an unmanned aerial vehicle, and optimizing shooting parameters by using a terrain adaptive parameter adjustment model to obtain a high-quality image; then, preprocessing and spectral decomposition and reconstruction are carried out on the image, and vegetation features are extracted and optimized; and then, fusing the processed image with the vegetation features to generate a high-resolution three-dimensional real scene model. And finally, planning, designing, implementing and monitoring an ecological restoration project based on the model. The invention also provides a system for realizing the method. Through the method, the scientificity and efficiency of an ecological restoration project can be improved, and powerful technical support is provided for mine ecological restoration.

Description

Technical Field

[0001] The present invention relates to the technical field of ecological restoration, and in particular to an ecological restoration project management method and system based on a three-dimensional real-scene model. Background Art

[0002] Mine ecological restoration is a complex and long-term process, which requires a comprehensive ecological assessment of the project area and the formulation of a scientific restoration plan based on this. However, traditional ground surveys and manual mapping methods are not only time-consuming and labor-intensive, but also difficult to fully and accurately reflect the complex terrain and vegetation conditions in the mining area. In recent years, with the rapid development of drone technology and remote sensing technology, it has become possible to use high-resolution aerial images for ecological monitoring and assessment. However, in the field of mine ecological restoration, how to effectively use these image data, extract key ecological information, and construct a three-dimensional model suitable for ecological restoration still faces many challenges.

[0003] First, the terrain in the mining area is complex and changeable, which places higher demands on the collection and processing of image data. Traditional image processing methods are often difficult to adapt to such complex terrain environments, resulting in poor image quality, which in turn affects subsequent ecological assessments and restoration plans.

[0004] Second, mine ecological restoration needs to pay attention to the recovery of vegetation, and the extraction of vegetation information is an important task in remote sensing image processing. However, due to the low vegetation coverage in the mining area and easy obscuration, the extraction of vegetation information becomes more difficult. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides an ecological restoration project management method and system based on a three-dimensional real-scene model.

[0006] On the one hand, an embodiment of the present invention provides an ecological restoration project management method based on a three-dimensional real scene model, the method comprising:

[0007] Acquire first mine image data and second mine image data collected by a drone;

[0008] determining an image processing strategy according to the terrain corresponding to the first mine image data, and preprocessing the first mine image data according to the determined image processing strategy to obtain a first image;

[0009] Performing spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image, calculating and analyzing a multi-scale vegetation index based on the reconstructed image, and optimizing the multi-scale vegetation index to obtain vegetation features;

[0010] fusing the first image and the vegetation feature to obtain a second image; and generating a three-dimensional real scene model based on the second image;

[0011] Carry out ecological restoration projects based on the generated 3D real-life model.

[0012] Preferably, before obtaining the first mine image data and the second mine image data collected by the drone, the method includes:

[0013] Acquire pre-shot image data;

[0014] Extracting terrain features from the pre-shot image data, inputting the terrain features into a terrain adaptive parameter adjustment model, and obtaining adaptive shooting parameters;

[0015] Shooting is performed according to the adaptive shooting parameters to obtain first mine image data and second mine image data.

[0016] Preferably, the terrain adaptive parameter adjustment model is:

[0017]

[0018] In the formula, Padapt(t) represents the adaptive shooting parameters at time t, T(t) represents the terrain features at time t, δT(t) represents the change in terrain features at time t, θ and ψ are parameters, γ represents the learning rate, r(τ, a(τ) represents the immediate reward after taking an action at time t, and the action is the shooting parameter.

[0019] Preferably, performing spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image includes:

[0020] Based on the wavelet transform method and independent component analysis method, the preprocessed second mine image is decomposed into multiple spectral components; the decomposed spectral components are subjected to feature extraction to obtain spectral features;

[0021] The spectral features are optimized and reconstructed using a reconstruction algorithm based on sparse representation and deep learning to obtain a reconstructed image.

[0022] Preferably, the multi-scale vegetation index is optimized to obtain vegetation features, including: optimizing the multi-scale vegetation index in combination with texture features and shape features, specifically:

[0023]

[0024] Where, MVI opt represents the optimized multi-scale vegetation index, VI s Indicates the sth vegetation index, TF hhth texture feature, TF uth shape feature, β s , α h , γ u They represent the corresponding weights respectively, GT represents the ground truth, R(β,α,γ) represents the regularization term, and λ represents the regularization parameter.

[0025] Preferably, constructing a three-dimensional real scene model based on the second image further includes:

[0026] Using a pre-built super-resolution model, the resolution of the second image is improved to obtain a target image;

[0027] A three-dimensional real scene model is constructed based on the target image.

[0028] On the other hand, an embodiment of the present invention provides an ecological restoration project management system based on a three-dimensional real scene model, the system comprising:

[0029] An acquisition module, used to acquire first mine image data and second mine image data collected by a drone;

[0030] A first processing module, configured to determine an image processing strategy according to the terrain corresponding to the first mine image data, and pre-process the first mine image data according to the determined image processing strategy to obtain a first image;

[0031] a second processing module, configured to perform spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image, calculate and analyze a multi-scale vegetation index based on the reconstructed image, optimize the multi-scale vegetation index in combination with texture features and shape features, and obtain vegetation features;

[0032] A three-dimensional model building module, which fuses the first image and the vegetation feature to obtain a second image, and generates a three-dimensional real scene model based on the second image;

[0033] The ecological restoration module is used to carry out ecological restoration projects based on the generated three-dimensional real-scene model.

[0034] The beneficial effects of the present invention are embodied in that the present invention provides an ecological restoration project management method and system based on a three-dimensional real-life model. The method first collects image data of the mining area through an unmanned aerial vehicle, and uses a terrain adaptive parameter adjustment model to optimize the shooting parameters to obtain high-quality images. Next, the image is preprocessed and spectrally decomposed and reconstructed to extract and optimize vegetation features. Then, the processed image is fused with the vegetation features to generate a high-resolution three-dimensional real-life model. Finally, the ecological restoration project is planned, designed, implemented and monitored based on the model. The present invention also provides a system for implementing the method. Through the present invention, the scientificity and efficiency of ecological restoration projects can be improved, providing strong technical support for mine ecological restoration. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0036] Figure 1 A flowchart of an ecological restoration project management method based on a three-dimensional real scene model is provided for an embodiment of the present invention;

[0037] Figure 2 A structural schematic diagram of an ecological restoration project management system based on a three-dimensional real-scene model is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention.

[0039] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0040] Example 1

[0041] like Figure 1 As shown, an embodiment of the present invention provides an ecological restoration project management method based on a three-dimensional real scene model, the method comprising:

[0042] Step 1: Acquire first mine image data and second mine image data collected by a drone.

[0043] In an embodiment of the present invention, before obtaining the first mine image data and the second mine image data collected by the drone, the method includes: obtaining pre-shot image data; extracting terrain features from the pre-shot image data, inputting the terrain features into a terrain adaptive parameter adjustment model, and obtaining adaptive shooting parameters; shooting according to the adaptive shooting parameters to obtain the first mine image data and the second mine image data.

[0044] The drone is equipped with a high-definition camera and a spectral camera. It takes off and shoots along a preset flight route. During the flight, the drone adjusts shooting parameters such as shutter speed, aperture, ISO, etc. in real time according to the terrain adaptive parameter adjustment model to ensure that high-quality images can be captured under different terrain and lighting conditions. After shooting, the drone transmits the image data to the ground station for subsequent processing.

[0045] Specifically, we use the high-precision positioning data of the UAV and the images taken by the camera, combined with Kriging interpolation, radial basis function network and deep learning network to build a high-precision terrain model, and accurately predict the terrain changes during the flight of the UAV through the real-time terrain prediction algorithm. We design a nonlinear terrain response function, dynamically adjust the camera's tilt angle, focal length and other shooting parameters according to the results of the terrain model prediction, and introduce an adaptive adjustment algorithm based on reinforcement learning or deep reinforcement learning to adjust and optimize the response function online. This is achieved specifically through the following formula:

[0046]

[0047] Wherein, Padapt(t) represents the adaptive shooting parameter at time t, T(t) represents the terrain feature at time t, δT(t) represents the change of terrain feature at time t, θ and ψ are parameters, γ represents the learning rate, r(τ, a(τ) represents the immediate reward after taking an action at time t, and the action is the shooting parameter.

[0048] It should also be noted that the embodiment of the present invention introduces an immediate function to determine the immediate reward. In the adaptive adjustment process, the reward function is used to evaluate the immediate reward obtained after taking a specific action at a certain moment. This reward can be based on a variety of factors, such as the image quality after the shooting parameters are adjusted, the degree of matching of terrain features, or other indicators related to the task objectives. A reward weighted model is trained based on these indicators to obtain an immediate reward function. The reward function provided in this embodiment comprehensively considers multiple factors such as image quality, degree of matching of terrain features and cost. By adjusting the weight coefficient, the trade-off and optimization of different factors can be achieved. In practical applications, the reward function can be designed and adjusted according to specific task requirements and conditions.

[0049] Traditional methods usually use fixed shooting parameters, which may cause overexposure, dark or blurred images under certain terrain and lighting conditions, and cannot adjust the shooting parameters in real time according to terrain characteristics, so high-quality image data may not be captured. The embodiments of the present invention can dynamically adjust the shooting parameters according to different terrain and lighting conditions, improve the quality and accuracy of the image, reduce image quality problems caused by improper shooting parameters, and improve the reliability of subsequent image processing and analysis.

[0050] Step 2: determining an image processing strategy according to the terrain corresponding to the first mine image data, and preprocessing the first mine image data according to the determined image processing strategy to obtain a first image.

[0051] Specifically, a preliminary analysis is performed on the first mine image data to extract terrain features, such as height, slope, vegetation coverage, etc., and a suitable image processing strategy is selected according to the terrain features, such as denoising, enhancement, correction, etc. The first mine image data is preprocessed, and the selected image processing strategy is applied to obtain an optimized first image.

[0052] Traditional methods usually use a unified image processing strategy, which may not be optimized for different terrains, and thus cannot effectively remove noise from the image or enhance the details of the image, resulting in reduced accuracy of subsequent analysis and processing. However, the embodiments of the present invention can select the most appropriate image processing strategy according to the terrain characteristics, improve image quality and usability, enhance the details and clarity of the image, and improve the accuracy of subsequent vegetation index calculation and analysis.

[0053] Step 3: spectrally decompose and reconstruct the second mine image to obtain a reconstructed image, calculate and analyze a multi-scale vegetation index based on the reconstructed image, and optimize the multi-scale vegetation index to obtain vegetation features.

[0054] In an embodiment of the present invention, the second mine image is spectrally decomposed and reconstructed to obtain a reconstructed image, including: decomposing the preprocessed second mine image into multiple spectral components based on wavelet transform method and independent component analysis method; extracting features of the decomposed spectral components to obtain spectral features; and optimizing and reconstructing the spectral features using a reconstruction algorithm based on sparse representation and deep learning to obtain a reconstructed image.

[0055] Spectral decomposition can separate the spectral information in the image, which is convenient for subsequent analysis and processing; through feature extraction and optimized reconstruction, spectral information can be further extracted and utilized to improve the accuracy and reliability of vegetation index.

[0056] In an embodiment of the present invention, the multi-scale vegetation index is optimized to obtain vegetation features, including: optimizing the multi-scale vegetation index in combination with texture features and shape features, specifically:

[0057]

[0058] Where, MVI opt represents the optimized multi-scale vegetation index, VI s Indicates the sth vegetation index, TF h hth texture feature, TF uth shape feature, β s , α h , γ u They represent the corresponding weights respectively, GT represents the ground truth, R(β,α,γ) represents the regularization term, and λ represents the regularization parameter.

[0059] Traditional methods cannot process complex spectral data, resulting in reduced reliability of subsequent analysis. However, the present invention can make full use of spectral information, improve the accuracy and reliability of vegetation index, enhance the extraction and analysis capabilities of vegetation information, and provide more accurate vegetation data support for ecological restoration projects.

[0060] Step 4: Fusing the first image and the vegetation features to obtain a second image; and generating a three-dimensional real scene model based on the second image.

[0061] In an embodiment of the present invention, constructing a three-dimensional real scene model based on the second image also includes: using a pre-constructed super-resolution model to improve the resolution of the second image to obtain a target image; and constructing a three-dimensional real scene model based on the target image.

[0062] Specifically, the second image is processed using a pre-trained super-resolution model. This model can be based on deep learning technology, such as a convolutional neural network (CNN), which can learn the mapping relationship between low-resolution images and high-resolution images. Through the super-resolution model, the resolution of the second image is significantly improved, the details are clearer, and higher-quality input is provided for subsequent 3D modeling. Using the target image with improved resolution as input, a 3D real scene model is constructed using 3D modeling software or algorithms (such as 3D Max, Maya, or a 3D reconstruction algorithm based on computer vision).

[0063] This model not only contains terrain information, but also integrates vegetation characteristics, providing a detailed and accurate virtual environment for ecological restoration projects.

[0064] Step 5: Conduct ecological restoration projects based on the generated 3D real-life model.

[0065] Specifically, the terrain data in the three-dimensional real-life model is used to conduct elevation analysis, slope and aspect analysis, etc., to identify areas that require key restoration and potential safety hazards. The vegetation feature information in the model is used to evaluate the vegetation coverage and species diversity in different areas, providing a basis for vegetation restoration and planting plans. The results of terrain analysis and vegetation distribution assessment are combined with project goals and budgets to develop a scientific and reasonable ecological restoration plan.

[0066] In summary, the embodiment of the present invention provides an ecological restoration project management method based on a three-dimensional real-life model. The method first collects image data of the mining area through a drone, and uses a terrain adaptive parameter adjustment model to optimize the shooting parameters to obtain high-quality images. Next, the image is preprocessed and spectrally decomposed and reconstructed to extract and optimize vegetation features. Then, the processed image is fused with the vegetation features to generate a high-resolution three-dimensional real-life model. Finally, the ecological restoration project is planned, designed, implemented and monitored based on the model.

[0067] Example 2

[0068] like Figure 2 As shown, based on the same inventive concept, an embodiment of the present invention provides an ecological restoration project management system based on a three-dimensional real scene model, the system comprising:

[0069] An acquisition module, used to acquire first mine image data and second mine image data collected by a drone;

[0070] A first processing module, configured to determine an image processing strategy according to the terrain corresponding to the first mine image data, and pre-process the first mine image data according to the determined image processing strategy to obtain a first image;

[0071] a second processing module, configured to perform spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image, calculate and analyze a multi-scale vegetation index based on the reconstructed image, optimize the multi-scale vegetation index in combination with texture features and shape features, and obtain vegetation features;

[0072] A three-dimensional model building module, which fuses the first image and the vegetation feature to obtain a second image, and generates a three-dimensional real scene model based on the second image;

[0073] The ecological restoration module is used to carry out ecological restoration projects based on the generated three-dimensional real-scene model.

[0074] It should be understood that an ecological restoration project management system based on a three-dimensional real-scene model provided in an embodiment of the present invention and an ecological restoration project management method based on a three-dimensional real-scene model provided in the above embodiment are based on the same inventive concept. For more specific working principles of each module in the embodiment of the present invention, please refer to the above embodiment and will not be repeated in the embodiment of the present invention.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. The ecological restoration project management method based on three-dimensional real scene model is characterized by: include: Acquire first mine image data and second mine image data collected by a drone; determining an image processing strategy according to the terrain corresponding to the first mine image data, and preprocessing the first mine image data according to the determined image processing strategy to obtain a first image; Performing spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image, calculating and analyzing a multi-scale vegetation index based on the reconstructed image, and optimizing the multi-scale vegetation index to obtain vegetation features; fusing the first image and the vegetation feature to obtain a second image; generating a three-dimensional real scene model based on the second image; Carry out ecological restoration projects based on the generated 3D real-life model.

2. The method according to claim 1, characterized in that Before obtaining the first mine image data and the second mine image data collected by the drone, the following steps are included: Acquire pre-shot image data; Extracting terrain features from the pre-shot image data, inputting the terrain features into a terrain adaptive parameter adjustment model, and obtaining adaptive shooting parameters; Shooting is performed according to the adaptive shooting parameters to obtain first mine image data and second mine image data.

3. The method according to claim 2, characterized in that The terrain adaptive parameter adjustment model is: Wherein, Padapt(t) represents the adaptive shooting parameter at time t, T(t) represents the terrain feature at time t, δT(t) represents the change of terrain feature at time t, θ and ψ are parameters, γ represents the learning rate, r(τ, a(τ) represents the immediate reward after taking an action at time t, and the action is the shooting parameter.

4. The method according to claim 1, characterized in that: Performing spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image includes: Based on the wavelet transform method and independent component analysis method, the preprocessed second mine image is decomposed into multiple spectral components; the decomposed spectral components are subjected to feature extraction to obtain spectral features; The spectral features are optimized and reconstructed using a reconstruction algorithm based on sparse representation and deep learning to obtain a reconstructed image.

5. The method according to claim 1, characterized in that Optimizing the multi-scale vegetation index to obtain vegetation features includes: optimizing the multi-scale vegetation index by combining texture features and shape features, specifically: Where, MVI opt represents the optimized multi-scale vegetation index, VI s Indicates the sth vegetation index, TF h The hth texture feature, TF, the uth shape feature, β s , α h , γ u They represent the corresponding weights respectively, GT represents the ground truth, R(β,α,γ) represents the regularization term, and λ represents the regularization parameter.

6. The method according to claim 1, characterized in that Building a three-dimensional real scene model based on the second image also includes: Using a pre-built super-resolution model, the resolution of the second image is improved to obtain a target image; A three-dimensional real scene model is constructed based on the target image.

7. The ecological restoration project management system based on the three-dimensional real scene model is characterized by: include: An acquisition module, used to acquire first mine image data and second mine image data collected by a drone; A first processing module, configured to determine an image processing strategy according to the terrain corresponding to the first mine image data, and pre-process the first mine image data according to the determined image processing strategy to obtain a first image; a second processing module, configured to perform spectral decomposition and reconstruction on the second mine image to obtain a reconstructed image, calculate and analyze a multi-scale vegetation index based on the reconstructed image, optimize the multi-scale vegetation index in combination with texture features and shape features, and obtain vegetation features; A three-dimensional model building module, which fuses the first image and the vegetation feature to obtain a second image, and generates a three-dimensional real scene model based on the second image; The ecological restoration module is used to carry out ecological restoration projects based on the generated three-dimensional real-scene model.

Images