A method and system for constructing a three-dimensional real-scene model of a water area breach

Abstract

The present invention relates to the field of image processing technology, and particularly to a method and system for constructing a three-dimensional model of the actual scene of a water area breach. The method includes the following steps: obtaining the position data of control points in the breach area; verifying the surveying and mapping basis for the position data of control points in the breach area to generate the reference data of control points in the breach area; performing three-dimensional coordinate transformation through the reference data of control points in the breach area to generate the plane coordinate data of the breach area; receiving the satellite signals of the drone based on the plane coordinate data of the breach area to obtain the standard satellite signals of the drone; collecting the aerial images of the breach area based on the standard satellite signals of the drone to obtain the aerial images of the breach area. The present invention improves the accuracy and timeliness of the three-dimensional model through precise data acquisition, efficient drone image acquisition and processing, dynamic image utilization, and optimized point cloud model construction and texture mapping.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a method and system for constructing a three-dimensional real-scene model of a water area breach. Background Art

[0002] Early water area management mainly relied on traditional measurement techniques such as hydrological observation and topographic surveying, which could not effectively reflect complex water flow and topographic changes. With the development of computer technology, especially the introduction of GIS (Geographic Information System) and remote sensing technology, the acquisition and processing of water area information have become more efficient. Entering the 21st century, three-dimensional modeling technology has gradually matured, especially the application of emerging technologies such as lidar (LiDAR) and unmanned aerial vehicles (UAVs), which has significantly improved the accuracy and efficiency of the water area breach area. LiDAR can quickly obtain high-resolution topographic data, while UAVs can cover large areas in a short time and capture subtle topographic changes. The combination of these technologies provides a rich data basis for the construction of a three-dimensional model of a water area breach. In recent years, the progress of deep learning and artificial intelligence technologies has made automation and intelligence in the data processing and model construction processes possible. However, currently, due to unclear boundary recognition of breach images in traditional model construction and the lack of change perception in the breach area in traditional static models, the accuracy and timeliness of three-dimensional models are relatively low. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for constructing a three-dimensional real-scene model of a water area breach to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for constructing a three-dimensional real-scene model of a water area breach, the method includes the following steps:

[0005] Step S1: Obtain the position data of control points in the breach area; confirm the surveying and mapping basis for the position data of control points in the breach area to generate the reference data of control points in the breach area; perform three-dimensional coordinate transformation through the reference data of control points in the breach area to generate the plane coordinate data of the breach area; receive the satellite signal of the standard UAV based on the plane coordinate data of the breach area.

[0006] Step S2: Collect aerial images of the breach area based on the standard UAV satellite signal to obtain the aerial images of the breach area; perform HSV color space conversion on the aerial images of the breach area to generate the color space conversion images of the breach area; use the color space images of the breach area to divide the sea-land areas of the standard aerial images of the breach area to generate the sea area map of the breach and the land area map of the breach; perform hierarchical dynamic image processing on the sea area map of the breach and the land area map of the breach to generate the dynamic sea area map of the breach and the dynamic land area map of the breach.

[0007] Step S3: Stitch the dynamic map of the offshore area of the breach and the dynamic map of the onshore area of the breach to generate a dynamic map of the water area of the breach; perform image texture fusion on the dynamic map of the water area of the breach to generate a texture map of the water area of the breach;

[0008] Step S4: Construct an initial point cloud model based on the optimized data of the water area breach point cloud to generate an initial point cloud model of the water area breach; perform UV unwrapping on the initial point cloud model of the water area breach, and map the texture map of the water area breach to the unwrapped initial point cloud model of the water area breach for texture mapping to generate a texture mapping model of the water area breach; perform model rendering on the texture mapping model of the water area breach to generate a three-dimensional real scene model of the water area breach.

[0009] The present invention obtains the position data of control points in the breach area through measurement and satellite positioning technology. These control points are the basis for subsequent analysis and modeling. The obtained control point data is confirmed for the surveying and mapping datum to ensure the accuracy and reliability of the data, and the control point datum data of the breach area is generated. Three-dimensional coordinate transformation is performed using the datum data to generate the plane coordinate data of the breach area. This step converts the three-dimensional data into a two-dimensional plane representation for convenient subsequent processing. Based on the plane coordinate data, satellite signal reception of the unmanned aerial vehicle (UAV) is performed to obtain the standard UAV satellite signal, providing navigation support for subsequent image acquisition. This step ensures the accuracy of the control points, laying the foundation for subsequent image acquisition and data processing, thereby improving the accuracy and efficiency of the entire process. Aerial photography is carried out using the standard UAV satellite signal to obtain the aerial images of the breach area. The obtained aerial images are subjected to HSV color space transformation to better analyze the color information, and the color space transformation images of the breach area are generated. The aerial images are analyzed using the color space images to divide the marine area map and the land area map of the breach. Layered dynamic image processing is performed on the marine area map and the land area map to generate the dynamic map of the marine area of the breach and the dynamic map of the land area of the breach. This step realizes the accurate identification and classification of the breach area through detailed color analysis and dynamic processing, providing rich visual information for subsequent image stitching and texture processing. The dynamic map of the marine area of the breach and the dynamic map of the land area of the breach are stitched to generate the dynamic map of the water area breach area. Texture fusion processing is performed on the generated dynamic map to obtain the texture map of the water area breach, enhancing the visual effect. Through image stitching and texture fusion, a high-quality dynamic map of the water area breach area is generated, facilitating subsequent three-dimensional model construction and analysis, and improving the realism and visualization effect of the model. An initial point cloud model is constructed based on the optimized point cloud data of the water area breach. The initial point cloud model is subjected to UV unwrapping, and the texture map of the water area breach is mapped onto the unwrapped model to generate the texture mapping model of the water area breach. The texture mapping model is rendered to generate the real-scene three-dimensional model of the water area breach. This process makes the model not only realistic in shape but also more vivid in visual effect by mapping the texture onto the three-dimensional model, providing high-quality three-dimensional visual data for subsequent analysis, decision-making, and display. Therefore, the present invention improves the accuracy and timeliness of the three-dimensional model through accurate data acquisition, efficient UAV image acquisition and processing, utilization of dynamic images, and optimized point cloud model construction and texture mapping.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Obtain the position data of control points in the breach area;

[0012] Step S12: Confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate transformation through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area;

[0013] Step S13: Based on the plane coordinate data of the breach area, deploy the UAV flight, and receive the UAV satellite signal through the GNSS receiver on the UAV to obtain the UAV satellite signal;

[0014] Step S14: Preprocess the UAV satellite signal to obtain the standard UAV satellite signal.

[0015] Through the acquisition of the control point position data and the confirmation of the surveying and mapping basis, the present invention ensures the accurate spatial position of the breach area and provides a reliable reference coordinate system. This is crucial for subsequent three-dimensional reconstruction and other engineering applications. Through coordinate transformation, the three-dimensional information of the breach area is planarized, simplifying the complexity of UAV flight deployment, improving the usability and accuracy of spatial data. Based on the plane coordinate data for UAV flight deployment, the flight path and task execution efficiency of the UAV are optimized, manual operation errors are reduced, and the coverage and accuracy of data collection are improved. Using the satellite signal obtained by the UAV through the GNSS receiver can achieve higher-precision positioning and navigation, enhancing the monitoring ability of the breach area. Preprocessing the collected UAV satellite signal lays a solid foundation for subsequent analysis and applications.

[0016] Preferably, step S12 includes the following steps:

[0017] Step S121: Adjust the reference parameters for the position data of the control points in the breach area to generate the plane coordinate reference parameter data and the elevation reference parameter data; establish the reference framework of the control points in the area according to the plane coordinate reference parameter data and the elevation reference parameter data to generate the reference framework data of the control points in the breach area;

[0018] Step S122: Based on the reference framework data of the control points in the breach area, correct the elevation difference of the position data of the control points in the breach area to generate the elevation correction data of the control points in the breach area;

[0019] Step S123: Perform plane coordinate transformation on the position data of the control points in the breach area through the three-degree zone coordinate transformation rule to generate the reference data of the control points in the breach area; set the initial parameters of the three-dimensional coordinates according to the reference data of the control points in the breach area to obtain the initial parameter data of the three-dimensional coordinates, where the initial parameter setting of the three-dimensional coordinates includes the translation amount, the rotation amount, and the scaling factor;

[0020] Step S124: Perform seven-parameter solution on the position data of the control points in the breach area according to the initial three-dimensional coordinate parameter data to obtain the seven parameters for three-dimensional coordinate transformation, where the seven-parameter solution includes three translation solutions, three rotation solutions, and a scaling solution; perform three-dimensional coordinate transformation on the elevation correction data of the control points in the breach area through the seven parameters for three-dimensional coordinate transformation, so as to generate the plane coordinate data of the breach area.

[0021] Through step S121 of the present invention, the plane and elevation reference parameters of the position data of the control points in the breach area are adjusted, a reference framework for the regional control points is established, ensuring the accuracy and consistency of the coordinate system, and providing a reliable spatial reference for subsequent operations. The elevation correction in step S122 can eliminate the elevation differences between the control points in the breach area, further improve the data accuracy, and avoid the influence of elevation errors on the coordinate accuracy, which is particularly important in the case of complex terrain. Step S123 accurately converts the position data of the control points in the breach area into plane coordinates based on the three-degree zone coordinate transformation rule. At the same time, the initial parameters of the three-dimensional coordinates are set, and through the translation amount, rotation amount, and scaling factor, the precise calibration of the three-dimensional spatial position is ensured. Through the seven-parameter solution of three translations, three rotations, and scaling, a higher-precision coordinate transformation can be achieved in the three-dimensional space, making the positioning of the control point data in the breach area more accurate in space. This solution method can minimize the errors in the conversion process. Through the transformation of the elevation correction data by the seven parameters, the finally generated plane coordinate data takes into account both the elevation differences and performs accurate three-dimensional coordinate solution, thus providing a high-precision and reliable three-dimensional coordinate system for the spatial data of the breach area.

[0022] Preferably, step S2 includes the following steps:

[0023] Step S21: Collect aerial images of the breach area by an unmanned aerial vehicle based on the standard unmanned aerial vehicle satellite signal to obtain the aerial images of the breach area; perform image preprocessing on the aerial images of the breach area to generate standard aerial images of the breach area, where the image preprocessing includes image filtering, image brightness enhancement, and image geometric transformation;

[0024] Step S22: Perform HSV color space conversion on the standard aerial images of the breach area to generate the color space conversion images of the breach area; use the color space images of the breach area to perform segmentation on the standard aerial images of the breach area, so as to generate a rough segmentation area map of the breach;

[0025] Step S23: Confirm the segmentation boundary of the rough segmentation area map of the breach to obtain the boundary data of the segmentation area of the breach; perform sea-land area division on the rough segmentation area map of the breach based on the boundary data of the segmentation area of the breach, so as to generate a sea area map of the breach and a land area map of the breach;

[0026] Step S24: Perform layered dynamic image processing on the offshore area map and onshore area map of the breach to generate an offshore area dynamic map and an onshore area dynamic map of the breach.

[0027] In the present invention, aerial photography is carried out based on standardized drone satellite signals, ensuring the accuracy and consistency of image acquisition in the breach area. During the image preprocessing process, filtering, brightness enhancement, and geometric transformation are adopted, which can effectively improve the image quality and make subsequent analysis more accurate. By converting the standard aerial photography image into the HSV color space, the ability to identify color differences in the breach area is enhanced, which helps to more accurately extract information on the breach area. At the same time, rough segmentation of the breach area is performed using the color space image, providing a fast and effective preliminary area division. Step S23 further confirms the boundary of the roughly segmented area, improving the segmentation accuracy. Subsequently, based on the segmentation boundary data, the sea-land area division is carried out, clearly separating the offshore part and onshore part of the breach area, providing more detailed area information for breach monitoring. By performing layered dynamic image processing on the offshore and onshore areas, dynamic monitoring images are generated, which can reflect the changes in the breach area in real time. This dynamic image processing method can effectively capture the dynamic characteristics of the breach area, helping to timely identify and evaluate the development and potential risks of the breach.

[0028] Preferably, the segmentation of the standard aerial photography image of the breach area using the color space image of the breach area includes:

[0029] Extract the color brightness characteristics of the color space image of the breach area to obtain the hue of the pixels in the breach area, the brightness of the pixels in the breach area, and the saturation of the breach area; calculate the hue change amplitude of the pixels in the breach area to obtain the hue change amplitude data of the pixels in the breach area; divide the hue of the pixels in the breach area through the hue change amplitude data of the pixels in the breach area to generate a gently changing hue and a fluctuating hue in the breach area.

[0030] Perform fuzzy segmentation of the offshore area on the standard aerial photography image of the breach area according to the gently changing hue and saturation of the breach area to generate a fuzzy segmentation image of the offshore area; perform fuzzy segmentation of the onshore area on the standard aerial photography image of the breach area according to the fluctuating hue and pixel brightness of the breach area to generate a fuzzy segmentation image of the onshore area.

[0031] Perform pixel color clustering on the fuzzy segmentation image of the offshore area and the fuzzy segmentation image of the onshore area to generate a rough segmentation area map of the breach.

[0032] By extracting the hue, brightness, and saturation features of the pixels in the breach area, the present invention can accurately capture the changes in different color features within the breach area. This process can effectively distinguish the spectral features of the ocean and land areas, providing a solid foundation for subsequent segmentation. By calculating the amplitude of hue change, the hue change in the breach area can be refined into two categories: gentle change and fluctuating change. The gentle change in hue mainly corresponds to the ocean area, while the fluctuating change in hue reflects the complexity of the land area. This division improves the accuracy of segmentation. According to the hue change characteristics, fuzzy segmentation of the ocean and land areas is performed, enabling effective identification of the regional boundaries in a complex natural environment. Even in the presence of transitional or confusing areas of color and brightness, the accuracy of segmentation can be ensured. The gentle hue change in the ocean area is combined with saturation, and the fluctuating hue change in the land area is combined with brightness, achieving fuzzy segmentation of different areas. Pixel color clustering is performed on the ocean and land area images after fuzzy segmentation, further enhancing the fineness of segmentation. This process not only integrates the color information of adjacent pixels but also effectively removes fine color noise through the clustering algorithm, improving the accuracy and clarity of the segmented image of the breach area. The finally generated rough segmentation area map of the breach can accurately reflect the main features of the breach area, providing accurate regional division data for subsequent disaster analysis, emergency response, and terrain modeling.

[0033] Preferably, the confirmation of the segmentation boundary for the rough segmentation area map of the breach includes:

[0034] Performing local feature texture calculation on the rough segmentation area map of the breach to obtain the pixel feature values of the breach segmentation area;

[0035] Among them, the formula for local feature texture calculation is as follows:

[0036]

[0037] Denoted as the coordinate feature value of the pixel point In the rough segmentation area map of the breach, Denoted as the abscissa in the rough segmentation area map of the breach, Denoted as the ordinate in the rough segmentation area map of the breach, Denoted as the value of the central pixel, Denoted as the value of the neighborhood pixel, Denoted as the number of neighborhood pixels, Denoted as the index;

[0038] Compare the pixel feature values of the breach segmentation area with the preset standard texture feature threshold. When the pixel feature value of the breach segmentation area is greater than the preset standard texture feature threshold, mark the pixels in the corresponding rough segmentation area map of the breach as pixels with complex texture features; when the pixel feature value of the breach segmentation area is less than the preset standard texture feature threshold, mark the pixels in the corresponding rough segmentation area map of the breach as pixels with simple texture features;

[0039] Based on the pixels with simple texture features and the pixels with complex texture features, confirm the segmentation boundary of the rough segmentation area map of the breach, and mark the pixels corresponding to the pixel feature value of the breach segmentation area equal to the preset standard texture feature threshold as segmentation boundary pixels, so as to obtain the boundary data of the breach segmentation area.

[0040] The present invention uses the local feature texture calculation formula , which can effectively extract the local texture features of each pixel in the rough segmentation area map of the breach. This formula captures the local texture differences by comparing the gray values of the central pixel and the neighborhood pixels to generate texture feature values. This calculation method ensures the accurate extraction of texture features and lays a foundation for the confirmation of the segmentation boundary. By comparing the calculated pixel feature values of the breach segmentation area with the preset standard texture feature threshold, the pixels can be divided into pixels with simple texture features and pixels with complex texture features. This differentiation method helps to further accurately identify different structures within the breach area, especially the texture features of complex terrains, which is helpful for better determining the boundary. When the pixel feature value is equal to the preset standard texture feature threshold, these pixels are automatically marked as segmentation boundary pixels. This step realizes the automatic detection of the segmentation boundary, avoids the uncertainty of manual judgment, and at the same time ensures the consistency and accuracy of the segmentation result. Based on the comparison of the pixels with simple and complex texture features, the segmentation boundary of the breach area can be more carefully confirmed, especially in areas with significant texture changes. In this way, the determination of the boundary not only depends on large-scale color or brightness changes, but also takes into account local texture differences, making the boundary line more accurate. Through the fine calculation of pixel feature values and threshold judgment, accurate boundary data of the breach segmentation area can be generated, providing high-precision boundary information for subsequent breach monitoring, analysis and emergency management. This boundary confirmation method based on feature values effectively improves the reliability of segmentation and is applicable to automatic breach analysis in complex environments.

[0041] Preferably, the hierarchical dynamic image processing of the offshore area map of the breach and the onshore area map of the breach includes the following steps:

[0042] Continuously capture frame images of the sea area of the breach to obtain consecutive frame images of the sea area of the breach; set a reference frame for the consecutive frame images of the sea area of the breach to obtain a reference frame image of the sea area of the breach; perform image difference calculation on the consecutive frame images of the sea area of the breach based on the reference frame image of the sea area of the breach to obtain a difference image of the sea area of the breach;

[0043] Perform maximum inter-class variance calculation on the difference image of the sea area of the breach to obtain variance data of the sea area of the breach, and use the variance data of the sea area of the breach as a threshold to mark the fluctuating area image of the difference image of the sea area of the breach to obtain a dynamic image of the sea area of the breach;

[0044] Capture multi-view images of the land area of the breach to obtain multi-view images of the land area of the breach; separate complex background layers from the multi-view images of the land area of the breach to generate a separated image of the land area of the breach, where the separated image of the land area of the breach includes a terrain separated image of the land area of the breach and a non-terrain separated image of the land area of the breach.

[0045] Through continuous frame capture and difference image calculation, the present invention can accurately capture the dynamic changes in the sea area of the breach, especially the water body fluctuation area. Through maximum inter-class variance calculation, the fluctuating area can be effectively screened out, making the dynamic monitoring of the sea area more accurate. For the multi-view image capture and processing of the land area, through the separation of complex background layers, a clear distinction can be made between the terrain and non-terrain areas, thereby improving the accuracy of image analysis, especially in complex environments. This hierarchical dynamic image processing method can adopt corresponding algorithms for image processing according to the different characteristics of the sea and land areas, reducing the need for human intervention, and improving the automation and intelligence levels of image processing. By performing hierarchical processing on the sea and land areas, more comprehensive monitoring data of the breach area can be provided, helping analysts accurately understand the actual situation of the breach area, especially the terrain and dynamic changes, and contributing to the response to emergencies.

[0046] Preferably, separating complex background layers from the multi-view images of the land area of the breach includes:

[0047] Extract the top-down view image of the land area of the breach from the multi-view images of the land area of the breach to obtain the top-down view image of the land area of the breach; confirm the image center point of the top-down view image of the land area of the breach to obtain the center point of the top-down view image of the land area of the breach;

[0048] Screen the multi-view images of the land area of the breach according to the center point of the top-down view image of the land area of the breach to generate an inclined view image of the land area of the breach; perform spatial height structure annotation on the top-down view image of the land area of the breach through the inclined view image of the land area of the breach to generate spatial height structure area data;

[0049] Based on the data of the spatial height structure area, the multi-view images of the land area of the breach are separated into background terrain layers to generate the separated images of the land area of the breach, where the separated images of the land area of the breach include the terrain separated images of the land area of the breach and the non-terrain separated images of the land area of the breach.

[0050] Through the combination of images from the top-down view and the oblique view, the present invention can perform multi-angle analysis on the land area of the breach, especially the spatial height structure annotation. This can better determine the terrain changes within the area and improve the accuracy of the terrain structure. The separation of the terrain and non-terrain layers based on the data of the spatial height structure area can effectively separate the terrain features and complex environments in the background, so that complex background information such as vegetation and buildings in the non-terrain area will not interfere with the terrain analysis, thereby improving the resolvability of the image. The use of multi-view image processing means, especially the combination of the oblique view and the top-down view, helps to capture terrain information more comprehensively, avoid blind spots under a single view, and provide a more three-dimensional environmental perception effect. Through automatic confirmation of the image center point and view angle screening, the background layer separation can be quickly carried out in a large amount of data, reducing manual intervention, improving the image processing efficiency, and accelerating the assessment speed of the disaster area. The finally generated terrain and non-terrain separated images can provide more refined data support for subsequent geographical analysis, disaster assessment and emergency response, and provide a reliable basis for the dynamic changes and feature recognition of the breach area.

[0051] Preferably, step S3 includes the following steps:

[0052] Step S31: Stitch the dynamic image of the sea area of the breach and the dynamic image of the land area of the breach to generate a dynamic image of the water area of the breach;

[0053] Step S32: Sparsify the point cloud of the dynamic image of the water area of the breach to generate a point cloud image of the water area of the breach; perform point cloud downsampling on the point cloud image of the water area of the breach to generate optimized data of the point cloud of the water area of the breach;

[0054] Step S33: Calculate the pixel normal vector of the point cloud image of the water area of the breach to obtain the pixel normal vector of the point cloud of the water area of the breach;

[0055] Step S34: Perform texture fusion on the point cloud image of the water area of the breach according to the pixel normal vector of the point cloud of the water area of the breach to generate a texture map of the water area of the breach.

[0056] By stitching the dynamic images of the offshore and onshore areas of the breach, a complete dynamic map of the water area of the breach is generated, which can intuitively display the overall dynamic changes in the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparse point clouding the dynamic map of the water area of the breach, the two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of the point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional space analysis. The calculation of the normal vector of the pixels of the breach point cloud in the water area helps to extract the surface geometric features in the point cloud. Especially in the complex water environment of the breach, geometric information such as the slope and surface change of the breach area can be accurately analyzed. This provides accurate geometric data support for subsequent structural analysis and modeling. By performing texture fusion on the normal vector of the pixels of the breach point cloud in the water area, a high-quality texture image is generated, which can more delicately represent the surface features of the water area and the breach area. Such a texture image provides rich texture details for subsequent three-dimensional visualization and rendering, improving the visual performance effect. Through the complete processes of image stitching, point clouding, normal vector calculation, and texture fusion, the three-dimensional morphological structure of the water area of the breach can be comprehensively and accurately presented. This provides a scientific basis and data support for disaster assessment, rescue plan formulation, and water area management.

[0057] In this specification, a construction system for a three-dimensional real-scene model of a water area breach is provided, which is used to execute the above-mentioned method for constructing a three-dimensional real-scene model of a water area breach. The construction system for the three-dimensional real-scene model of the water area breach includes:

[0058] A plane conversion module, which is used to obtain the position data of the control points in the breach area; confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate conversion through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area; and receive the standard UAV satellite signal based on the plane coordinate data of the breach area.

[0059] A regional subdivision module, which is used to collect the aerial images of the breach area based on the standard UAV satellite signal to obtain the aerial images of the breach area; perform HSV color space conversion on the aerial images of the breach area to generate the color space conversion images of the breach area; use the color space images of the breach area to divide the sea-land areas of the standard aerial images of the breach area to generate the offshore area map and the onshore area map of the breach; and perform layered dynamic image processing on the offshore area map and the onshore area map of the breach to generate the dynamic map of the offshore area of the breach and the dynamic map of the onshore area of the breach.

[0060] A point cloud conversion module is used to splice the dynamic map of the offshore area of the breach and the dynamic map of the onshore area of the breach to generate a dynamic map of the water area of the breach; perform image texture fusion on the dynamic map of the water area of the breach to generate a texture map of the water area of the breach.

[0061] A model construction module is used to construct an initial point cloud model based on the optimized data of the water area breach point cloud to generate an initial point cloud model of the water area breach; perform UV unfolding on the initial point cloud model of the water area breach, and map the texture map of the water area breach to the unfolded initial point cloud model of the water area breach for texture mapping to generate a texture mapping model of the water area breach; perform model rendering on the texture mapping model of the water area breach to generate a real-scene three-dimensional model of the water area breach.

[0062] The beneficial effects of the present invention are as follows: By splicing the dynamic images of the offshore area and the onshore area of the breach, a complete dynamic map of the water area of the breach is generated, which can intuitively display the overall dynamic changes in the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparse point clouding the dynamic map of the water area of the breach, a two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of the point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional space analysis. The calculation of the normal vector of the water area breach point cloud pixels helps to extract the surface geometric features in the point cloud. Especially in the complex water area environment of the breach, geometric information such as the slope and surface change of the breach area can be accurately analyzed. This provides accurate geometric data support for subsequent structural analysis and modeling. By performing texture fusion on the normal vector of the water area breach point cloud pixels, a high-quality texture image is generated, which can more delicately represent the surface features of the water area and the breach area. Such a texture image provides rich texture details for subsequent three-dimensional visualization and rendering, improving the visual performance effect. Through the complete process of image splicing, point clouding, normal vector calculation, and texture fusion, the three-dimensional morphological structure of the water area breach can be comprehensively and accurately displayed. This provides a scientific basis and data support for disaster assessment, rescue plan formulation, and water area management. Therefore, the present invention improves the accuracy and timeliness of the three-dimensional model through accurate data acquisition, efficient UAV image acquisition and processing, utilization of dynamic images, and optimized point cloud model construction and texture mapping. Description of the Drawings

[0063] Figure 1 It is a schematic diagram of the step flow of a method for constructing a real-scene three-dimensional model of a water area breach;

[0064] Figure 2 For Figure 1 a detailed implementation step flow diagram of step S2 in;

[0065] Figure 3 For Figure 1Schematic diagram of the detailed implementation steps of step S3 in

[0066] The realization, functional features and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Specific implementation manner

[0067] The technical method of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0068] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0069] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0070] To achieve the above object, please refer to Figures 1 to 3 , a method for constructing a three-dimensional real scene model of a water area breach, the method comprising the following steps:

[0071] Step S1: Obtain the position data of the control points in the breach area; confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate conversion through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area; receive the drone satellite signal based on the plane coordinate data of the breach area to obtain the standard drone satellite signal;

[0072] Step S2: Collect drone aerial images based on standard drone satellite signals to obtain aerial images of the breach area; perform HSV color space conversion on the aerial images of the breach area to generate color space conversion images of the breach area; use the color space images of the breach area to divide the aerial images of the standard breach area into sea and land areas of the breach, thereby generating a sea area map of the breach and a land area map of the breach; perform hierarchical dynamic image processing on the sea area map of the breach and the land area map of the breach to generate a dynamic map of the sea area of the breach and a dynamic map of the land area of the breach;

[0073] Step S3: Stitch the dynamic map of the sea area of the breach and the dynamic map of the land area of the breach to generate a dynamic map of the water area of the breach; perform image texture fusion on the dynamic map of the water area of the breach to generate a texture map of the water area of the breach;

[0074] Step S4: Construct an initial point cloud model based on the optimized data of the water area breach point cloud to generate an initial point cloud model of the water area breach; perform UV unwrapping on the initial point cloud model of the water area breach, and map the texture map of the water area of the breach to the unwrapped initial point cloud model of the water area breach for texture mapping to generate a texture mapping model of the water area breach; perform model rendering on the texture mapping model of the water area breach to generate a three-dimensional real scene model of the water area breach.

[0075] The present invention obtains the position data of control points in the breach area through measurement and satellite positioning technology. These control points are the basis for subsequent analysis and modeling. The obtained control point data is confirmed for the mapping datum to ensure the accuracy and reliability of the data, and the control point datum data of the breach area is generated. Three-dimensional coordinate transformation is performed using the datum data to generate the plane coordinate data of the breach area. This step converts three-dimensional data into a two-dimensional plane representation for convenient subsequent processing. Based on the plane coordinate data, the satellite signal reception of the unmanned aerial vehicle (UAV) is carried out to obtain the standard UAV satellite signal, providing navigation support for subsequent image acquisition. This step ensures the accuracy of the control points, laying the foundation for subsequent image acquisition and data processing, thereby improving the accuracy and efficiency of the entire process. The standard UAV satellite signal is used for aerial photography to obtain the aerial images of the breach area. The obtained aerial images are subjected to HSV color space transformation to better analyze the color information, generating the color space transformation images of the breach area. The aerial images are analyzed using the color space images to divide the marine area map and the land area map of the breach. Layered dynamic image processing is performed on the marine area map and the land area map to generate the dynamic map of the marine area of the breach and the dynamic map of the land area of the breach. This step realizes the accurate identification and classification of the breach area through detailed color analysis and dynamic processing, providing rich visual information for subsequent image stitching and texture processing. The dynamic map of the marine area of the breach and the dynamic map of the land area of the breach are stitched to generate the dynamic map of the water area breach area. Texture fusion processing is performed on the generated dynamic map to obtain the texture map of the water area breach, enhancing the visual effect. Through image stitching and texture fusion, a high-quality dynamic map of the water area breach area is generated, facilitating subsequent three-dimensional model construction and analysis, and improving the realism and visualization effect of the model. An initial point cloud model is constructed based on the optimized point cloud data of the water area breach. The initial point cloud model is subjected to UV unwrapping, and the texture map of the water area breach is mapped onto the unwrapped model to generate the texture mapping model of the water area breach. The texture mapping model is rendered to generate the real-scene three-dimensional model of the water area breach. This process makes the model not only real in shape but also more vivid in visual effect by mapping the texture onto the three-dimensional model, providing high-quality three-dimensional visual data for subsequent analysis, decision-making, and display. Therefore, the present invention improves the accuracy and timeliness of the three-dimensional model through accurate data acquisition, efficient UAV image acquisition and processing, dynamic image utilization, and optimized point cloud model construction and texture mapping.

[0076] In the embodiment of the present invention, with reference to Figure 1 as described, it is a schematic diagram of the step flow of a method for constructing a real-scene three-dimensional model of a water area breach of the present invention. In this example, the method for constructing a real-scene three-dimensional model of a water area breach includes the following steps:

[0077] Step S1: Obtain the position data of the control points in the breach area; confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate transformation through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area; receive the UAV satellite signal based on the plane coordinate data of the breach area to obtain the standard UAV satellite signal;

[0078] In the embodiment of the present invention, a plurality of control points are arranged in the breach area by using a GPS positioning device to ensure the accuracy of the position data of each control point. Record the three-dimensional coordinate data (longitude, latitude, altitude) of each control point for subsequent processing. Compare the obtained position data of the control points with the existing surveying and mapping basis (such as the national surveying and mapping coordinate system). Use a ground surveying instrument (such as a total station) for on-site confirmation to ensure the accurate position of the control points, generate the reference data of the control points in the breach area, and record the standard coordinates and relevant information of each control point. Based on the reference data of the control points in the breach area, transform the original three-dimensional coordinates to generate plane coordinate data, and the obtained plane coordinate data of the breach area will be used for subsequent UAV operations. Deploy the UAV at the confirmed plane coordinate position of the breach area and ensure that the GNSS (Global Navigation Satellite System) module of the UAV is in good working condition. Receive the UAV satellite signal to ensure that the UAV obtains the standard UAV satellite signal and complete the position calibration. Record the signal strength, the number of satellites and their positions received by the UAV to ensure the accuracy of subsequent data collection.

[0079] Step S2: Collect the aerial images of the breach area based on the standard UAV satellite signal to obtain the aerial images of the breach area; perform HSV color space conversion on the aerial images of the breach area to generate the color space conversion images of the breach area; use the color space images of the breach area to divide the aerial images of the standard breach area into sea and land areas of the breach, so as to generate the sea area map of the breach and the land area map of the breach; perform layered dynamic image processing on the sea area map of the breach and the land area map of the breach to generate the dynamic sea area map of the breach and the dynamic land area map of the breach;

[0080] In the embodiments of the present invention, an unmanned aerial vehicle (UAV) is utilized to conduct aerial photography based on the standard UAV satellite signals obtained in step S1. Ensure that the UAV flies stably within the breach area and conducts multi-angle aerial photography according to the preset flight path. Record the time, location, and altitude information of the aerial photography images for subsequent analysis. Perform HSV (hue, saturation, value) color space conversion on the obtained aerial photography images of the breach area to better analyze the color features in the images. The color space conversion images of the breach area obtained using color space conversion formulas (such as RGB conversion formula and HSV conversion formula) will be used for subsequent area division. Utilize the color information in the color space images of the breach area to conduct sea-land area division on the standard aerial photography images of the breach area. Set thresholds according to the characteristics of hue (H) and saturation (S) to divide the pixels in the image into sea area and land area, generating two binary images: the breach sea area map and the breach land area map. Conduct continuous frame image acquisition on the breach sea area map to obtain the continuous frame images of the breach sea area. Conduct multi-view image acquisition on the breach land area map to obtain the multi-view images of the breach land area. Conduct dynamic change analysis on these two types of images respectively: use the image difference method to calculate the dynamic changes in the sea area and mark the fluctuating areas. Conduct complex background separation on the multi-view images to extract terrain and non-terrain features.

[0081] Step S3: Stitch the dynamic images of the breach sea area and the dynamic images of the breach land area to generate a dynamic image of the water area breach area; perform image texture fusion on the dynamic image of the water area breach area to generate a texture map of the water area breach area;

[0082] In the embodiments of the present invention, by ensuring that the dynamic map of the offshore area of the breach and the dynamic map of the onshore area of the breach are sorted in time series and their timestamps and perspectives are aligned, seamless stitching is achieved. In each frame of the image, a feature point detection algorithm (such as SIFT or ORB) is used to identify key feature points. The feature points are matched according to their positions and descriptors to determine the overlapping area. The transformation matrix of the stitched image is calculated, and the wrong matches are removed through the RANSAC algorithm to ensure accurate image alignment. The image is geometrically transformed using the transformation matrix to seamlessly stitch the dynamic map of the offshore area and the dynamic map of the onshore area. An image blending technique (such as multi-band fusion) is used to process the overlapping area to avoid obvious visible boundaries, ensure an overall natural transition, and output the final stitched image to form a dynamic map of the water area of the breach. Texture features are extracted from each frame of the stitched dynamic map of the water area of the breach. Techniques such as local binary pattern (LBP) and Gabor filters can be used to obtain texture features. The extracted texture features are applied to the original image, and the feature performance is enhanced by adjusting texture mapping parameters (such as intensity and scale). A specific image fusion algorithm (such as weighted average, image pyramid fusion, etc.) is selected to fuse the texture features with the original image. This process ensures the retention of key features and the enhancement of details. Color correction is performed on the fused image to enhance the contrast, making the image more natural and clear in details. Gaussian blur or median filtering is used for denoising to improve the image quality, remove existing artifacts and noise, and output the final water area texture map, which will be used for further analysis or visualization.

[0083] Step S4: Based on the optimized data of the water area breach point cloud, an initial point cloud model is constructed to generate an initial point cloud model of the water area breach; the initial point cloud model of the water area breach is UV-unwrapped, and the water area texture map is mapped onto the unwrapped initial point cloud model of the water area breach for texture mapping to generate a texture mapping model of the water area breach; the texture mapping model of the water area breach is model-rendered to generate a 3D real-scene model of the water area breach.

[0084] In the embodiments of the present invention, based on the optimized data of the water area breach point cloud, it is ensured that the data undergoes downsampling and denoising processes to obtain high-quality point cloud data. A point cloud reconstruction algorithm (such as Poisson reconstruction or surface reconstruction method) is used to generate an initial three-dimensional point cloud model of the water area breach. Ensure that the model surface is smooth to avoid defects caused by noise or discontinuities, and output the completed initial point cloud model of the water area breach. Perform UV unwrapping on the initial point cloud model to generate a UV coordinate map. The UV unwrapping can be completed through specific UV unwrapping algorithms (such as the cutting method, spherical projection, etc.) to reduce distortion during texture mapping. Ensure a reasonable UV layout and avoid overlapping areas to facilitate subsequent texture mapping. Select the previously generated texture map of the water area breach to ensure that the texture quality and resolution are suitable for application to the three-dimensional model. Map the texture map of the water area breach to the initial point cloud model of the water area breach after UV unwrapping. Use texture coordinates to accurately apply the texture to the model surface. Perform texture adjustment to ensure that the texture matches the geometric shape of the model to avoid stretching or compression phenomena, and output the texture-mapped model of the water area breach after texture mapping processing. Select a suitable rendering engine (such as Blender, Unity, Unreal Engine, etc.) and import the texture-mapped model into it. Configure lighting, material properties, and other rendering parameters to ensure that the generated real-scene model conforms to the actual effect. Perform model rendering to generate the final real-scene three-dimensional model of the water area breach. Different rendering modes (such as real-time rendering, ray tracing, etc.) can be selected during the rendering process to improve the visual effect. Perform post-processing on the rendering result, applying high dynamic range imaging (HDR) and color correction to enhance the visual quality. Save the rendered three-dimensional model to ensure that it can be used for subsequent display and analysis.

[0085] Preferably, step S1 includes the following steps:

[0086] Step S11: Obtain the position data of the control points in the breach area;

[0087] Step S12: Confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate conversion through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area;

[0088] Step S13: Based on the plane coordinate data of the breach area, deploy the drone flight, and receive the drone satellite signal through the GNSS receiver on the drone to obtain the drone satellite signal;

[0089] Step S14: Perform signal preprocessing on the drone satellite signal to obtain the standard drone satellite signal.

[0090] In the embodiments of the present invention, multiple control points are arranged in the breach area by selecting appropriate tools (such as GPS receivers, total stations), and their geographical coordinate (longitude, latitude, elevation) information is recorded. Ensure that the selected control points are representative and evenly distributed for subsequent data processing and analysis. Using the local surveying and mapping datum system, the position data of the control points obtained is verified for accuracy to ensure compliance with the standards. Record the verification results, including information such as the number, coordinates, and accuracy of the reference points, and generate the control point reference data for the breach area. Use specific coordinate transformation algorithms (such as geodetic coordinate transformation method, plane coordinate transformation method) to perform three-dimensional coordinate transformation on the control point reference data to obtain plane coordinate data (usually represented by projected coordinates). Ensure that the terrain characteristics and related errors of the area are considered during the coordinate transformation process to improve the accuracy of the data. According to the plane coordinate data of the control points, plan the flight path of the unmanned aerial vehicle, including flight altitude, speed, flight line spacing, etc., to ensure coverage of the entire breach area. Use the unmanned aerial vehicle aerial photography software to set up the flight mission to ensure the safety and efficiency of the unmanned aerial vehicle flight. Install a GNSS receiver on the unmanned aerial vehicle to ensure its good satellite signal reception ability. Start the unmanned aerial vehicle and begin to fly. The GNSS receiver receives satellite signals in real time and records the movement trajectory of the unmanned aerial vehicle. During the flight of the unmanned aerial vehicle, the GNSS receiver continuously records the received satellite signals, and the data includes information such as timestamp, satellite ID, and pseudorange. Preprocess the collected satellite signals, including denoising, filtering, data synchronization, etc., to improve the signal quality. Use standardized methods (such as Kalman filtering, differential GPS, etc.) to process the signals and generate standard unmanned aerial vehicle satellite signals to ensure accurate and reliable data.

[0091] Preferably, step S12 includes the following steps:

[0092] Step S121: Adjust the reference parameters of the control point position data in the breach area to generate plane coordinate reference parameter data and elevation reference parameter data; establish a regional control point reference framework based on the plane coordinate reference parameter data and elevation reference parameter data to generate control point reference framework data for the breach area;

[0093] Step S122: Correct the elevation difference of the control point position data in the breach area based on the control point reference framework data for the breach area to generate elevation correction data for the control points in the breach area;

[0094] Step S123: Perform plane coordinate transformation on the control point position data in the breach area through the three-degree zone coordinate transformation rule to generate control point reference data for the breach area; set the initial three-dimensional coordinate parameters according to the control point reference data for the breach area to obtain initial three-dimensional coordinate parameter data, where the initial three-dimensional coordinate setting includes translation amount, rotation amount, and scaling factor;

[0095] Step S124: Perform seven-parameter solution on the position data of the control points in the breach area according to the initial three-dimensional coordinate parameter data to obtain the seven parameters for three-dimensional coordinate transformation, where the seven-parameter solution includes three translation solutions, three rotation solutions, and a scaling solution; perform three-dimensional coordinate transformation on the elevation correction data of the control points in the breach area through the seven parameters for three-dimensional coordinate transformation, so as to generate the plane coordinate data of the breach area.

[0096] In the embodiment of the present invention, by analyzing the position data of the control points in the breach area, the longitude, latitude, and elevation information thereof is extracted. Benchmark parameters are set, including the reference ellipsoid, projection method, regional coordinate system, etc., to generate the plane coordinate benchmark parameter data and the elevation benchmark parameter data. According to the plane coordinate benchmark parameter data and the elevation benchmark parameter data, a regional control point benchmark framework is constructed, including the selected control point coordinates and their mutual relationships, to generate the breach area control point benchmark framework data, and all the control point information within the benchmark framework is recorded, providing a basis for subsequent coordinate transformation and correction. Based on the breach area control point benchmark framework data, the elevation differences of each control point are calculated. By applying corresponding correction models (such as terrain correction, meteorological correction, etc.), the elevation correction values of each control point are obtained. The elevation difference correction is applied to the original control point elevation data to generate the elevation correction data of the control points in the breach area, ensuring the accuracy of the elevation data. According to the three-degree zone coordinate transformation rule, a specific mathematical model is used to perform plane coordinate transformation on the control point position data to generate the breach area control point benchmark data, and the transformed plane coordinate information is recorded. Based on the generated control point benchmark data, initial three-dimensional coordinate parameters are set, including: translation amount: the displacement amount of each coordinate axis. Rotation amount: the rotation angle around each coordinate axis. Scaling factor: the scaling ratio of the coordinate data, to obtain the initial three-dimensional coordinate parameter data, preparing for the subsequent seven-parameter solution. According to the initial three-dimensional coordinate parameter data, perform seven-parameter solution on the position data of the control points in the breach area, including: three translation solutions: calculate the translation amounts along the X, Y, and Z directions. Three rotation solutions: calculate the rotation angles around the X, Y, and Z axes. Scaling solution: calculate the scaling factor of the coordinate data, generate the seven parameters for three-dimensional coordinate transformation, and record the values of each parameter. Use the obtained seven parameters to perform three-dimensional coordinate transformation on the elevation correction data of the control points in the breach area to generate the plane coordinate data of the breach area. Ensure that the differences in coordinate systems are considered during the transformation process, and an algorithm is adopted to improve the accuracy and reliability of the transformation.

[0097] As an example of the present invention, refer to Figure 2 shown, in this example, the step S2 includes:

[0098] Step S21: Collect UAV aerial images based on standard UAV satellite signals to obtain aerial images of the breach area; perform image preprocessing on the aerial images of the breach area to generate standard aerial images of the breach area, where image preprocessing includes image filtering, image brightness enhancement, and image geometric transformation;

[0099] Step S22: Perform HSV color space conversion on the standard aerial images of the breach area to generate color space conversion images of the breach area; use the color space images of the breach area to segment the standard aerial images of the breach area, thereby generating a rough segmentation area map of the breach;

[0100] Step S23: Confirm the segmentation boundary of the rough segmentation area map of the breach to obtain the boundary data of the breach segmentation area; based on the boundary data of the breach segmentation area, divide the rough segmentation area map of the breach into sea and land areas, thereby generating a sea area map of the breach and a land area map of the breach;

[0101] Step S24: Perform hierarchical dynamic image processing on the sea area map of the breach and the land area map of the breach to generate a dynamic sea area map of the breach and a dynamic land area map of the breach.

[0102] In the embodiments of the present invention, by planning an aerial photography route based on the standard drone satellite signal, it is ensured that the drone covers the entire breach area. The drone is started and flown at a fixed altitude, and aerial photography images of the breach area are obtained according to the set flight altitude and speed. Ensure that the position and attitude information of the drone are recorded during the aerial photography process for subsequent image registration. The obtained aerial photography images are subjected to image filtering processing to remove noise and enhance the image quality. Median filtering or Gaussian filtering and other methods can be used, and appropriate parameters are selected to achieve the best effect. Histogram equalization or gamma correction technology is used to improve the brightness and contrast of the images, making the features of the breach area more obvious. Image geometric transformation is performed, including rotation, scaling, and cropping, to ensure the spatial consistency of the images and the accuracy of the target area, generating standard aerial photography images of the breach area and providing a clear data basis for subsequent processing. The standard aerial photography images of the breach area are converted into the HSV color space for better color analysis. The H (hue), S (saturation), and V (brightness) components are extracted for subsequent processing. Using the hue and saturation information in the color space image of the breach area, the standard aerial photography images of the breach area are segmented. Threshold segmentation or clustering algorithms (such as the K-means algorithm) are applied to extract the rough segmentation areas related to the breach features, generating a rough segmentation area map of the breach and clearly identifying the target area. Edge detection is performed on the rough segmentation area map of the breach to confirm the contour of the segmentation area. The Canny edge detection algorithm or contour extraction technology can be used to obtain clear boundary data of the segmentation area. Based on the confirmed boundary data of the breach segmentation area, the sea and land areas are divided. According to the characteristic differences between water bodies and land, a sea area map of the breach and a land area map of the breach are generated. The division results are recorded to ensure the accuracy of the area division. The generated sea area map of the breach and land area map of the breach are subjected to layering processing to extract information at different levels. Dynamic image processing technology is applied to generate a dynamic map of the sea area of the breach and a dynamic map of the land area of the breach, showing the changes at different time periods, and analyzing and visually displaying the dynamic changes.

[0103] Preferably, using the color space image of the breach area to segment the standard aerial photography images of the breach area includes:

[0104] Extract the color brightness features of the color space image of the breach area to obtain the pixel hue of the breach area, the pixel brightness of the breach area, and the saturation of the breach area; calculate the hue change amplitude of the pixel hue of the breach area to obtain the hue change amplitude data of the pixel hue of the breach area; divide the pixel hue of the breach area through the hue change amplitude data of the pixel hue of the breach area to generate a gently changing hue of the breach area and a fluctuating changing hue of the breach area;

[0105] Perform fuzzy segmentation of the ocean area on the aerial image of the standard breach area according to the gently varying hue of the breach area and the saturation of the breach area to generate a fuzzy segmentation image of the ocean area; perform fuzzy segmentation of the land area on the aerial image of the standard breach area according to the fluctuating hue of the breach area and the pixel brightness of the breach area to generate a fuzzy segmentation image of the land area;

[0106] Perform pixel color clustering on the fuzzy segmentation image of the ocean area and the fuzzy segmentation image of the land area to generate a rough segmentation area map of the breach.

[0107] In the embodiment of the present invention, by processing the color space image of the breach area, the hue, brightness, and saturation information of each pixel is extracted to generate three data sets: Pixel hue: representing the color attribute of each pixel, expressed in degrees. Pixel brightness: representing the brightness of each pixel, ranging from 0 to 255. Saturation: representing the vividness of each pixel color, ranging from 0 to 255. Perform statistical analysis on the extracted pixel hue data to calculate the hue change amplitude of each pixel. Use the following formula: Hue change amplitude = max(H)−min(H), where H represents the pixel hue of the breach area, to generate the pixel hue change amplitude data of the breach area, reflecting the color stability of different areas. According to the hue change amplitude data, classify the pixel hues. If the hue change amplitude is less than the set threshold, it is marked as a gently varying hue. If the hue change amplitude is greater than or equal to the set threshold, it is marked as a fluctuating hue, generating two data sets: the gently varying hue of the breach area and the fluctuating hue of the breach area. Use the gently varying hue and saturation data to perform fuzzy segmentation of the ocean area on the aerial image of the standard breach area. Adopt a fuzzy logic method or region growing algorithm to extract the ocean area according to the characteristics of hue and saturation to generate a fuzzy segmentation image of the ocean area. The threshold method can be used to set a suitable hue and saturation range to extract the ocean area. According to the fluctuating hue and pixel brightness, perform fuzzy segmentation of the land area on the aerial image of the standard breach area. Adopt a similar fuzzy logic method or region growing algorithm, combined with the characteristics of pixel brightness and fluctuating hue for extraction to generate a fuzzy segmentation image of the land area. By setting the thresholds of brightness and fluctuating change, ensure the accurate extraction of the land area. Perform pixel color clustering on the generated fuzzy segmentation image of the ocean area and the fuzzy segmentation image of the land area. Apply the K-means clustering or other clustering algorithms to group similar pixels into one class according to the color characteristics to generate a rough segmentation area map of the breach, identifying the characteristics of the breach area and its corresponding color attributes.

[0108] Preferably, the confirmation of the segmentation boundary for the rough segmentation area map of the breach includes:

[0109] Perform local feature texture calculation on the rough segmentation area map of the breach to obtain the pixel feature values of the breach segmentation area;

[0110] Among them, the formula for calculating the local feature texture is as follows:

[0111]

[0112] Denoted as pixel points The coordinate eigenvalue in the rough segmentation area map of the breach, Denoted as the abscissa in the rough segmentation area map of the breach, Denoted as the ordinate in the rough segmentation area map of the breach, Denoted as the value of the central pixel, Denoted as the value of the neighborhood pixel, Denoted as the number of neighborhood pixels, Denoted as the index;

[0113] Compare the pixel eigenvalue of the breach segmentation area with the preset standard texture feature threshold. When the pixel eigenvalue of the breach segmentation area is greater than the preset standard texture feature threshold, the pixel in the corresponding rough segmentation area map of the breach is marked as a complex texture feature pixel; when the pixel eigenvalue of the breach segmentation area is less than the preset standard texture feature threshold, the pixel in the corresponding rough segmentation area map of the breach is marked as a simple texture feature pixel;

[0114] Based on the simple texture feature pixels and complex texture feature pixels, confirm the segmentation boundary of the rough segmentation area map of the breach, and mark the pixels corresponding to the pixel eigenvalue of the breach segmentation area equal to the preset standard texture feature threshold as segmentation boundary pixels, so as to obtain the boundary data of the breach segmentation area.

[0115] In the embodiment of the present invention, for each pixel point In the rough segmentation area map of the breach, calculate its local feature texture eigenvalue , using the following formula: , where: Denoted as pixel points The coordinate eigenvalue in the rough segmentation area map of the breach, Denoted as the abscissa in the rough segmentation area map of the breach, Denoted as the ordinate in the rough segmentation area map of the breach, Denoted as the value of the central pixel, Denoted as the value of the neighborhood pixel, Denoted as the number of neighborhood pixels, Denoted as the index; traverse each pixel point in the rough segmentation area map of the breach, extract the values of the central pixel and its neighborhood pixels in turn, and calculate the corresponding texture eigenvalue . The preset standard texture feature threshold is denoted as For each calculated pixel feature value : If , then mark this pixel as a complex texture feature pixel. If , then mark this pixel as a simple texture feature pixel. Based on the simple texture feature pixels and complex texture feature pixels, further confirm the segmentation boundary of the rough segmentation area map of the breach. Mark the pixels with the feature value equal to the preset standard texture feature threshold as segmentation boundary pixels, and generate breach segmentation area boundary data, which contains the coordinate information of all pixels marked as segmentation boundary pixels.

[0116] Preferably, the hierarchical dynamic image processing of the offshore area map of the breach and the onshore area map of the breach includes the following steps:

[0117] Continuously collect frame images of the offshore area of the breach to obtain a continuous frame image of the offshore area of the breach; set a reference frame for the continuous frame image of the offshore area of the breach to obtain a reference frame image of the offshore area of the breach; perform image differential calculation on the continuous frame image of the offshore area of the breach based on the reference frame image of the offshore area of the breach to obtain a differential image of the offshore area of the breach;

[0118] Perform maximum inter-class variance calculation on the differential image of the offshore area of the breach to obtain variance data of the offshore area of the breach, and use the variance data of the offshore area of the breach as a threshold to mark the fluctuating area image of the differential image of the offshore area of the breach to obtain a dynamic map of the offshore area of the breach;

[0119] Collect multi-view images of the onshore area of the breach to obtain multi-view images of the onshore area of the breach; separate the complex background layer of the multi-view images of the onshore area of the breach to generate a separated image of the onshore area of the breach, where the separated image of the onshore area of the breach includes a terrain separated image of the onshore area of the breach and a non-terrain separated image of the onshore area of the breach.

[0120] In an embodiment of the present invention, by using a drone or other acquisition device to continuously collect frame images of the offshore area of the breach, a series of sequential images are recorded to form a continuous frame image sequence of the offshore area of the breach. Select one frame from the continuous frame image sequence as the reference frame, and this frame can be the first frame in the time series or a frame under specific conditions. Name the selected reference frame as the "reference frame image of the offshore area of the breach". Perform pixel-level differential calculation on each frame of the image and the reference frame image: , where represents the value of the differential image at the coordinate , represents the pixel value of the current frame image, represents the pixel value of the reference frame image. Perform the following operations on the differential image Perform maximum inter-class variance calculation to obtain the variance data of the image. Usually, Otsu's method can be used to determine the threshold. Calculate the variance data of the sea area of the breach, denoted as . Take the variance data of the sea area of the breach as the threshold, and mark the fluctuating area image of the sea area of the breach. Set the marking rule: If , then mark this area as the fluctuating area, generate the dynamic image of the sea area of the breach, and display all fluctuating areas. Collect multi-perspective images of the land area of the breach from different angles and positions, record the images under different perspectives, and form a multi-perspective image sequence of the land area of the breach. Process the multi-perspective images to separate the complex background: Apply image separation algorithms (such as image segmentation, foreground and background separation) to extract the terrain and non-terrain parts, and generate two separated images: The terrain separation image of the land area of the breach only contains terrain features. The non-terrain separation image of the land area of the breach contains non-terrain features such as the background and objects.

[0121] Preferably, the separation of complex background layers for the multi-perspective images of the land area of the breach includes:

[0122] Extract the top-down perspective image of the land area of the breach from the multi-perspective images of the land area of the breach to obtain the top-down perspective image of the land area of the breach; Confirm the center point of the image of the top-down perspective of the land area of the breach to obtain the center point of the top-down image of the land area of the breach;

[0123] Screen the inclined perspective images of the land area of the breach according to the center point of the top-down perspective image of the land area of the breach to generate the inclined perspective images of the land area of the breach; Mark the spatial height structure of the top-down perspective image of the land area of the breach through the inclined perspective images of the land area of the breach to generate the spatial height structure area data;

[0124] Separate the background terrain layers for the multi-perspective images of the land area of the breach based on the spatial height structure area data to generate the separated images of the land area of the breach, where the separated images of the land area of the breach include the terrain separation image of the land area of the breach and the non-terrain separation image of the land area of the breach.

[0125] In the embodiment of the present invention, the top-down perspective image is extracted from the collected multi-perspective images. This process can judge whether the image belongs to the top-down perspective by setting an angle threshold, usually set to a perspective close to vertical (for example, the angle is less than a preset value), and the obtained image is called the "top-down perspective image of the land area of the breach". Analyze the top-down perspective image of the land area of the breach to identify the center point of the image. The method adopted can be to calculate the geometric center of the image, and the formula is as follows: ; where and are the abscissa and ordinate of the center point respectively, is the total number of valid pixels in the image, is the top-down view image. The confirmed center point is used as the "center point of the top-down view image of the breach land area". Based on the center point of the top-down view image, multi-view images of the breach land area are screened for oblique view images. Set the screening rules, and select eligible oblique view images according to the angular deviation from the top-down center point. The generated image set is called the "oblique view image of the breach land area". The top-down view image is spatially height-structured annotated by the oblique view image of the breach land area. Using stereo vision or depth estimation methods, spatial height information is extracted from the oblique view image to generate spatial height structure area data, which contains height information of different regions. Based on the spatial height structure area data, the background terrain layer of the multi-view images of the breach land area is separated. Use hierarchical segmentation algorithms (such as image thresholding, clustering algorithms, etc.) to separate terrain information from non-terrain information. Finally, a separated image of the breach land area is generated, which includes: The terrain-separated image of the breach land area focuses on terrain features. The non-terrain separated image of the breach land area contains background and other object information.

[0126] As an example of the present invention, refer to Figure 3 shown, in this example, step S3 includes:

[0127] Step S31: Stitch the dynamic image of the breach sea area and the dynamic image of the breach land area to generate a dynamic image of the water area breach area;

[0128] Step S32: Sparsify the dynamic image of the water area breach area to generate a point cloud image of the water area breach; Downsample the point cloud image of the water area breach to generate optimized data of the point cloud of the water area breach;

[0129] Step S33: Calculate the pixel normal vector of the point cloud image of the water area breach to obtain the pixel normal vector of the point cloud of the water area breach;

[0130] Step S34: Perform texture fusion on the point cloud image of the water area breach according to the pixel normal vector of the point cloud of the water area breach to generate a texture map of the water area breach.

[0131] In the embodiments of the present invention, by splicing the dynamic map of the offshore area of the breach and the dynamic map of the onshore area of the breach. The feature matching algorithm (such as SIFT or ORB) is used to identify similar feature points in the overlapping area to ensure seamless connection. The alignment of the image is adjusted through perspective transformation (Homography) to generate a complete "dynamic map of the water area breach area". The generated dynamic map of the water area breach area is sparsely point-clouded to convert the image into a point-cloud format. Each pixel is converted into three-dimensional coordinates (x, y, z), where the z value is usually determined by the pixel brightness or depth information. Next, a point-cloud downsampling algorithm (such as random sampling or Voxel Grid filter) is applied to reduce the density of the point-cloud data and generate "optimized data of the water area breach point cloud", thereby reducing the computational burden of subsequent processing. The normal vector of each pixel point in the water area breach point-cloud image is calculated. The neighborhood points (such as k-neighborhood or fixed-radius neighborhood) are used to estimate the normal vector, and the formula is as follows: ; where is the normal vector of point , are the coordinates of the neighborhood points, are the coordinates of the initial point, is the number of neighborhood points. Finally, the "normal vector of the water area breach point-cloud pixel" is obtained, providing direction information for each point. Texture fusion is performed on the point-cloud image according to the normal vector of the water area breach point-cloud pixel. The normal vector information is used to resample the point cloud, so as to map the texture information of the image onto the point cloud. An interpolation algorithm (such as bilinear interpolation or cubic interpolation) is used to integrate the texture information into the point-cloud data to generate a "texture map of the water area breach". Finally, ensure that the generated texture map matches the point cloud, providing a clear view for subsequent visualization and analysis.

[0132] In this specification, a construction system for a three-dimensional real-scene model of a water area breach is provided, which is used to execute the above-mentioned method for constructing a three-dimensional real-scene model of a water area breach. The construction system for a three-dimensional real-scene model of a water area breach includes:

[0133] A plane conversion module, which is used to obtain the position data of the control points in the breach area; confirm the surveying and mapping basis for the position data of the control points in the breach area to generate the reference data of the control points in the breach area; perform three-dimensional coordinate conversion through the reference data of the control points in the breach area to generate the plane coordinate data of the breach area; receive the standard UAV satellite signal based on the plane coordinate data of the breach area;

[0134] The area subdivision module is used to collect aerial images of the drone based on the standard drone satellite signal to obtain aerial images of the breach area; perform HSV color space conversion on the aerial images of the breach area to generate color space conversion images of the breach area; use the color space images of the breach area to divide the sea and land areas of the standard breach area aerial images, thereby generating a sea area map of the breach and a land area map of the breach; perform layered dynamic image processing on the sea area map of the breach and the land area map of the breach to generate a dynamic map of the sea area of the breach and a dynamic map of the land area of the breach;

[0135] The point cloud conversion module is used to splice the dynamic maps of the sea area of the breach and the land area of the breach to generate a dynamic map of the water area breach; perform image texture fusion on the dynamic map of the water area breach to generate a texture map of the water area breach;

[0136] The model construction module is used to construct an initial point cloud model based on the optimized data of the water area breach point cloud to generate an initial point cloud model of the water area breach; perform UV unfolding on the initial point cloud model of the water area breach, and map the texture map of the water area breach to the unfolded initial point cloud model of the water area breach for texture mapping to generate a texture mapping model of the water area breach; perform model rendering on the texture mapping model of the water area breach to generate a three-dimensional real scene model of the water area breach.

[0137] The beneficial effects of the present invention are as follows: By splicing the dynamic images of the sea area and the land area of the breach, a complete dynamic map of the water area breach is generated, which can intuitively display the overall dynamic changes in the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparse point clouding the dynamic map of the water area breach, the two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of the point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional space analysis. The calculation of the normal vector of the water area breach point cloud pixels helps to extract the surface geometric features in the point cloud. Especially in a complex water area environment of the breach, geometric information such as the slope and surface change of the breach area can be accurately analyzed. This provides accurate geometric data support for subsequent structural analysis and modeling. By performing texture fusion on the normal vector of the water area breach point cloud pixels, a high-quality texture image is generated, which can more delicately represent the surface features of the water area and the breach area, provides rich texture details for subsequent three-dimensional visualization and rendering, and improves the visual performance effect. Through the complete processes of image splicing, point clouding, normal vector calculation, and texture fusion, the three-dimensional morphological structure of the water area breach can be comprehensively and accurately displayed. This provides a scientific basis and data support for disaster assessment, rescue plan formulation, and water area management. Therefore, the present invention improves the accuracy and timeliness of the three-dimensional model through accurate data acquisition, efficient drone image acquisition and processing, utilization of dynamic images, and optimized point cloud model construction and texture mapping.

[0138] Therefore, from every perspective, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Thus, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed by the present invention.

[0139] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a real-scene three-dimensional model of a water breach, characterized in that: The following steps are involved: Step S1: Obtaining the position data of the control points in the breach area; Confirm the location data of the control points in the breach area with the surveying and mapping benchmark, and generate the benchmark data of the control points in the breach area; perform three-dimensional coordinate conversion on the benchmark data of the control points in the breach area, and generate the plane coordinate data of the breach area; Based on the plane coordinate data of the breach area, the UAV satellite signal is received to obtain the standard UAV satellite signal; Step S2: based on the standard UAV satellite signal, drone aerial image acquisition is performed to obtain an aerial image of the breach area; image preprocessing is performed on the aerial image of the breach area to generate a standard aerial image of the breach area; HSV color space conversion is performed on the standard aerial image of the breach area to generate a color space conversion image of the breach area; the standard aerial image of the breach area is divided into breach sea and land areas by using the breach area color space image, thereby generating a breach sea area map and a breach on land area map; layered dynamic image processing is performed on the breach sea area map and the breach on land area map to generate a breach sea area dynamic map and a breach on land area dynamic map; Step S3: stitching the dynamic image of the breach sea area and the dynamic image of the breach land area to generate a dynamic image of the water breach area; performing image texture fusion on the dynamic image of the water breach area to generate a water breach texture map; Step S3 includes the following steps: Step S31: stitching the dynamic image of the breached sea area and the dynamic image of the breached land area to generate a dynamic image of the breached water area; Step S32: sparse point clouding is performed on the dynamic image of the water breach area to generate a water breach point cloud image; point cloud downsampling is performed on the water breach point cloud image to generate water breach point cloud optimization data; Step S33: Calculating pixel normal vectors of the water breach point cloud image to obtain pixel normal vectors of the water breach point cloud; Step S34: performing texture fusion on the water breach point cloud image according to the water breach point cloud pixel normal vector to generate a water breach texture map; Step S4: construct an initial point cloud model based on the water breach point cloud optimization data to generate an initial point cloud model of the water breach; UV unfold the initial point cloud model of the water breach, and map the water breach texture map to the unfolded initial point cloud model of the water breach for texture mapping to generate a water breach texture mapping model; render the water breach texture mapping model to generate a real-life three-dimensional model of the water breach.

2. The method for constructing a real-life three-dimensional model of a water breach according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: obtaining the position data of the control points in the breach area; Step S12: Confirm the position data of the breach area control points with a surveying and mapping benchmark to generate the breach area control point benchmark data; perform three-dimensional coordinate conversion on the breach area control point benchmark data to generate the breach area plane coordinate data; Step S13: performing flight deployment of the UAV based on the plane coordinate data of the breach area, and receiving the UAV satellite signal through the GNSS receiver on the UAV, thereby obtaining the UAV satellite signal; Step S14: Preprocess the UAV satellite signal to obtain a standard UAV satellite signal.

3. The method for constructing a real-life three-dimensional model of a water breach according to claim 2, characterized in that: Step S12 includes the following steps: Step S121: adjusting the reference parameters of the breach area control point position data to generate plane coordinate reference parameter data and elevation reference parameter data; establishing the regional control point reference framework according to the plane coordinate reference parameter data and elevation reference parameter data to generate the breach area control point reference framework data; Step S122: performing elevation difference correction on the position data of the breach area control points based on the reference framework data of the breach area control points to generate elevation correction data of the breach area control points; Step S123: performing plane coordinate conversion on the position data of the breach area control points by using the three-dimensional zone coordinate conversion rule to generate the breach area control point reference data; performing three-dimensional coordinate initial parameter setting according to the breach area control point reference data to obtain the three-dimensional coordinate initial parameter data, wherein the three-dimensional coordinate initial parameter setting includes the translation amount, the rotation amount and the scaling factor; Step S124: Perform a seven-parameter solution on the position data of the control points in the breach area according to the three-dimensional coordinate initial parameter data to obtain seven parameters for three-dimensional coordinate transformation, wherein the seven-parameter solution includes three translation solutions, three rotation solutions and a scaling solution; perform a three-dimensional coordinate transformation on the elevation correction data of the control points in the breach area through the seven parameters for three-dimensional coordinate transformation to generate the plane coordinate data of the breach area.

4. The method for constructing a real-life three-dimensional model of a water breach according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: collecting drone aerial images based on standard drone satellite signals to obtain aerial images of the breach area; performing image preprocessing on the aerial images of the breach area to generate standard aerial images of the breach area, wherein the image preprocessing includes image filtering, image brightness enhancement and image geometric transformation; Step S22: performing HSV color space conversion on the standard breach area aerial image to generate a breach area color space conversion image; performing breach area segmentation on the standard breach area aerial image using the breach area color space image to generate a breach rough segmentation area map; Step S23: confirming the segmentation boundary of the breach rough segmentation region map to obtain breach segmentation region boundary data; dividing the breach rough segmentation region map into sea and land regions based on the breach segmentation region boundary data, thereby generating a breach sea region map and a breach land region map; Step S24: performing layered dynamic image processing on the breach offshore area map and the breach onshore area map to generate a breach offshore area dynamic map and a breach onshore area dynamic map.

5. The method for constructing a real-scene three-dimensional model of a water breach according to claim 4, characterized in that: Using the breach area color space image to segment the breach area of ​​the standard breach area aerial image includes: The color brightness feature of the color space image of the ulcer area is extracted to obtain the pixel hue, pixel brightness and saturation of the ulcer area; the hue change amplitude of the pixel hue in the ulcer area is calculated to obtain the pixel hue change amplitude data of the ulcer area; the hue of the pixel hue in the ulcer area is divided into hues according to the pixel hue change amplitude data of the ulcer area to generate the smoothly changing hue and the fluctuating changing hue of the ulcer area; According to the smoothly changing hue and saturation of the breach area, the standard breach area aerial image is subjected to fuzzy segmentation of the ocean area to generate a fuzzy segmented image of the ocean area; according to the fluctuating hue and pixel brightness of the breach area, the standard breach area aerial image is subjected to fuzzy segmentation of the land area to generate a fuzzy segmented image of the land area; The pixel color clustering is performed on the fuzzy segmentation images of the ocean area and the fuzzy segmentation images of the land area to generate a rough segmentation area map of the breach.

6. The method for constructing a real-scene three-dimensional model of a water breach according to claim 4, characterized in that: The segmentation boundary confirmation of the rough segmentation area map of the breach includes: The local feature texture is calculated on the rough segmentation area map of the breach to obtain the pixel feature value of the breach segmentation area; Among them, the formula for calculating local feature texture is as follows: Represented as pixels The pixel feature value in the rough segmentation area map of the breach, It is represented as the horizontal coordinate in the rough segmentation area diagram of the breach. It is represented as the ordinate in the rough segmentation area diagram of the breach. Represented as the value of the center pixel, Represented as the value of the neighborhood pixel, Expressed as the number of neighborhood pixels, Represented as an index; The pixel feature value of the ulcer segmentation area is compared with the preset standard texture feature threshold. When the pixel feature value of the ulcer segmentation area is greater than the preset standard texture feature threshold, the corresponding pixel in the ulcer rough segmentation area map is marked as a complex texture feature pixel; when the pixel feature value of the ulcer segmentation area is less than the preset standard texture feature threshold, the corresponding pixel in the ulcer rough segmentation area map is marked as a simple texture feature pixel; The segmentation boundary of the rupture rough segmentation area map is confirmed based on simple texture feature pixels and complex texture feature pixels, and the pixels corresponding to the pixel feature values ​​of the rupture segmentation area equal to the preset standard texture feature threshold are marked as segmentation boundary pixels, thereby obtaining the rupture segmentation area boundary data.

7. The method for constructing a real-scene three-dimensional model of a water breach according to claim 4, characterized in that: The layered dynamic image processing of the breach offshore area map and the breach onshore area map includes the following steps: Perform continuous frame image acquisition on the breach marine area map to obtain continuous frame images of the breach marine area; perform reference frame setting on the continuous frame images of the breach marine area to obtain reference frame images of the breach marine area; perform image difference calculation on the continuous frame images of the breach marine area based on the reference frame images of the breach marine area to obtain a differential image of the breach marine area; The difference image of the breached offshore area is subjected to maximum inter-class variance calculation to obtain variance data of the breached offshore area, and the variance data of the breached offshore area is used as a threshold to mark the fluctuation area image of the difference image of the breached offshore area to obtain a dynamic image of the breached offshore area; Multi-view image acquisition is performed on the breach land area map to obtain a multi-view image of the breach land area; complex background layers are separated from the multi-view image of the breach land area to generate a breach land area separation image, wherein the breach land area separation image includes a breach land area terrain separation image and a breach land area non-terrain separation image.

8. The method for constructing a real-scene three-dimensional model of a water breach according to claim 7, characterized in that: The complex background layer separation of multi-view images of the breach land area includes: Extract the top-view image from the multi-view images of the breach land area to obtain the top-view image of the breach land area; confirm the center point of the top-view image of the breach land area to obtain the center point of the top-view image of the breach land area; According to the center point of the overhead image of the breach land area, the multi-perspective images of the breach land area are screened from oblique perspectives to generate oblique perspective images of the breach land area; spatial height structure annotation is performed on the overhead perspective images of the breach land area through the oblique perspective images of the breach land area to generate spatial height structure area data; Based on the spatial height structure area data, the background terrain layer of the multi-view image of the breach land area is separated to generate the breach land area separation image, wherein the breach land area separation image includes the breach land area terrain separation image and the breach land area non-terrain separation image.

9. A system for constructing a real-life three-dimensional model of a water breach, characterized in that: The method for constructing a real-life three-dimensional model of a water breach according to claim 1 is used to construct a real-life three-dimensional model of a water breach, and the system comprises: The plane conversion module is used to obtain the position data of the control points in the breach area; to perform surveying and mapping benchmark confirmation on the position data of the control points in the breach area to generate the benchmark data of the control points in the breach area; to perform three-dimensional coordinate conversion through the benchmark data of the control points in the breach area to generate the plane coordinate data of the breach area; to receive the UAV satellite signal based on the plane coordinate data of the breach area to obtain the standard UAV satellite signal; The regional segmentation module is used to collect drone aerial images based on standard drone satellite signals to obtain aerial images of the breach area; perform image preprocessing on the aerial images of the breach area to generate standard aerial images of the breach area; perform HSV color space conversion on the standard aerial images of the breach area to generate a color space conversion image of the breach area; use the color space image of the breach area to divide the standard aerial images of the breach area into sea and land areas of the breach, thereby generating a breach sea area map and a breach land area map; perform layered dynamic image processing on the breach sea area map and the breach land area map to generate a breach sea area dynamic map and a breach land area dynamic map; The point cloud conversion module is used to perform image stitching on the dynamic image of the breach sea area and the dynamic image of the breach land area to generate a dynamic image of the water breach area; perform image texture fusion on the dynamic image of the water breach area to generate a water breach texture image; The model building module is used to construct an initial point cloud model based on the water breach point cloud optimization data to generate an initial point cloud model of the water breach; UV unfold the initial point cloud model of the water breach, and map the water breach texture map to the unfolded initial point cloud model of the water breach for texture mapping to generate a water breach texture mapping model; and render the water breach texture mapping model to generate a real-life three-dimensional model of the water breach.

Images