Method and system for constructing real-scene three-dimensional model of water area breach

Abstract

The invention relates to the technical field of image processing, in particular to a construction method and system of a water area breach live-action three-dimensional model. The method comprises the following steps: acquiring breach area control point position data; performing surveying and mapping reference confirmation on the breach area control point position data to generate breach area control point reference data; performing three-dimensional coordinate conversion through the breach area control point reference data so as to generate breach area plane coordinate data; unmanned aerial vehicle satellite signal receiving is carried out based on the breach area plane coordinate data, and a standard unmanned aerial vehicle satellite signal is obtained; and performing unmanned aerial vehicle aerial image acquisition based on the standard unmanned aerial vehicle satellite signal to obtain an aerial image of the breach area. Through accurate data acquisition, efficient unmanned aerial vehicle image acquisition and processing, dynamic image utilization and optimized point cloud model construction and texture mapping, the accuracy and timeliness of the three-dimensional model are improved.

Description

Technical Field

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

[0002] Early water management mainly relied on traditional surveying techniques, such as hydrological observation and topographic mapping, which could not effectively reflect complex water flow and terrain 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 information has become more efficient. Entering the 21st century, three-dimensional modeling technology has gradually matured, especially the application of emerging technologies such as laser radar (LiDAR) and unmanned aerial vehicles (UAV), which has significantly improved the accuracy and efficiency of water breach areas. LiDAR can quickly obtain high-resolution terrain data, while UAVs can cover large areas in a short time and capture subtle terrain changes. The combination of these technologies provides a rich data foundation for the construction of three-dimensional models of water breaches. In recent years, the advancement of deep learning and artificial intelligence technology has made automation and intelligence possible in the process of data processing and model construction. However, the current traditional model construction has unclear boundary recognition of breach images and lacks change perception of breach areas in traditional static models, resulting in low accuracy and timeliness of three-dimensional models. Summary of the invention

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

[0004] To achieve the above object, a method for constructing a real-life three-dimensional model of a water breach is provided, the method comprising the following steps: Step S1: Acquire the position data of the breach area control points; perform surveying and mapping benchmark confirmation on the position data of the breach area control points to generate the benchmark data of the breach area control points; perform three-dimensional coordinate conversion on the benchmark data of the breach area control points 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; Step S2: based on the standard UAV satellite signal, drone aerial image acquisition is performed to obtain an aerial image of the breach area; the aerial image of the breach area is converted into an HSV color space to generate a breach area color space conversion image; the standard breach area aerial image is divided into breach sea and land areas using the breach area color space image, thereby generating a breach sea area map and a breach land area map; the breach sea area map and the breach land area map are subjected to layered dynamic image processing to generate a breach sea area dynamic map and a breach land area dynamic map; Step S3: performing 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 breach water area; performing image texture fusion on the dynamic image of the breach water area to generate a texture image of the breach water area; 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.

[0005] The present invention obtains the control point position data of 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 by the surveying and mapping benchmark to ensure the accuracy and reliability of the data and generate the control point benchmark data of the breach area. The benchmark data is used to perform three-dimensional coordinate conversion to generate the plane coordinate data of the breach area. This step is to convert the three-dimensional data into a two-dimensional plane representation to facilitate subsequent processing. The satellite signal of the unmanned aerial vehicle is received based on the plane coordinate data to obtain the standard unmanned aerial vehicle satellite signal to provide navigation support for subsequent image acquisition. This step ensures the accuracy of the control point and lays the foundation for subsequent image acquisition and data processing, thereby improving the accuracy and efficiency of the entire process. Aerial photography is performed using standard unmanned aerial vehicle satellite signals to obtain aerial images of the breach area. The acquired aerial images are converted into HSV color space to better analyze the color information and generate color space conversion images of the breach area. The aerial images are analyzed using color space images to divide the sea area map and the land area map of the breach. The sea area map and the land area map are subjected to layered dynamic image processing to generate dynamic images of the sea area and the land area of ​​the breach. This step achieves accurate identification and classification of the breach area through meticulous color analysis and dynamic processing, providing rich visual information for subsequent image stitching and texture processing. The dynamic images of the breach offshore area and the dynamic images of the land area are stitched to generate a dynamic image of the water breach area. The generated dynamic image is texture fused to obtain a texture map of the water breach to enhance the visual effect. Through image stitching and texture fusion, a high-quality dynamic image of the water breach area is generated, which is convenient for subsequent three-dimensional model construction and analysis, and improves the realism and visualization of the model. The initial point cloud model is constructed based on the point cloud optimization data of the water breach. The initial point cloud model is UV unfolded, and the water breach texture map is mapped to the unfolded model to generate a texture mapping model of the water breach. The texture mapping model is rendered to generate a real-life three-dimensional model of the water breach. This process makes the model not only realistic in shape but also more vivid in visual effect by mapping the texture to 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 precise data acquisition, efficient drone image collection and processing, dynamic image utilization, and optimized point cloud model construction and texture mapping.

[0006] Preferably, step S1 comprises 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.

[0007] The present invention ensures the accuracy of the spatial position of the breach area by acquiring control point position data and confirming the surveying and mapping benchmark, and provides a reliable reference coordinate system. This is crucial for subsequent three-dimensional reconstruction and other engineering applications. Through coordinate conversion, the three-dimensional information of the breach area is planarized, which simplifies the complexity of drone flight deployment and improves the usability and accuracy of spatial data. Drone flight deployment based on plane coordinate data optimizes the flight path and mission execution efficiency of the drone, reduces manual operation errors, and improves the coverage and accuracy of data collection. Using satellite signals obtained by drones using GNSS receivers can achieve higher-precision positioning and navigation, enhancing the monitoring capabilities of the breach area. Preprocessing the collected drone satellite signals lays a solid foundation for subsequent analysis and application.

[0008] Preferably, step S12 comprises 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.

[0009] The present invention adjusts the plane and elevation reference parameters of the control point position data of the breach area through step S121, establishes a reference framework of the regional control points, ensures the accuracy and consistency of the coordinate system, and provides a reliable spatial reference for subsequent operations. The elevation correction of step S122 can eliminate the elevation differences between the control points in the breach area, further improve the accuracy of the data, 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 control point position data of the breach area into plane coordinates based on the three-dimensional zone coordinate conversion rule. At the same time, the initial parameter setting of the three-dimensional coordinates ensures the accurate calibration of the three-dimensional space position through the translation amount, rotation amount and scaling factor. Through the seven-parameter solution of three translations, three rotations and scaling, a higher-precision coordinate conversion can be achieved in three-dimensional space, so that the control point data of the breach area can be more accurately positioned in space. This solution method can minimize the error in the conversion process. By converting the elevation correction data using seven parameters, the plane coordinate data finally generated not only takes the elevation difference into account, but also accurately calculates the three-dimensional coordinates, thus providing a high-precision and reliable three-dimensional coordinate system for the spatial data of the breach area.

[0010] Preferably, step S2 comprises 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.

[0011] The present invention ensures the accuracy and consistency of the image acquisition of the breach area by performing aerial photography based on standardized unmanned aerial vehicle satellite signals. In the image preprocessing process, filtering, brightness enhancement and geometric transformation are used to effectively improve the image quality and make the subsequent analysis more accurate. By converting the standard aerial image to the HSV color space, the color difference recognition ability of the breach area is enhanced, which helps to extract the information of the breach area more accurately. At the same time, the 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 rough segmentation area to improve the accuracy of the segmentation. Subsequently, the sea and land area is divided based on the segmentation boundary data, and the offshore part of the breach area is clearly distinguished from the onshore part, providing more detailed regional information for breach monitoring. By performing layered dynamic image processing on the offshore and onshore areas, a dynamic monitoring image is generated to 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 and help to timely identify and evaluate the development and potential risks of the breach.

[0012] Preferably, using the breach area color space image to segment the breach area of ​​the standard breach area aerial image comprises: 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.

[0013] The present invention can accurately capture the changes of different color features in the breach area by extracting features of the hue, brightness and saturation of the pixels in the breach area. This process can effectively distinguish the spectral characteristics of the ocean and land areas, providing a solid foundation for subsequent segmentation. By calculating the hue change amplitude, the hue change in the breach area can be refined into two categories: gentle change and fluctuating change. The gentle change hue mainly corresponds to the ocean area, while the fluctuating change hue reflects the complexity of the land area. This division improves the accuracy of segmentation. Fuzzy segmentation of the ocean and land areas is performed according to the hue change characteristics, so that the regional boundaries can be effectively identified in a complex natural environment, and the accuracy of segmentation can be guaranteed even if there are transitional or confusing areas of color and brightness. 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, so as to realize fuzzy segmentation of different areas. Pixel color clustering is performed on the ocean and land area images after fuzzy segmentation, which further enhances the meticulousness of segmentation. This process not only integrates the color information of adjacent pixels, but also effectively removes subtle color noise through the clustering algorithm, thereby improving the accuracy and clarity of the segmented image of the breach area. The resulting rough segmentation area map of the breach can accurately reflect the main characteristics of the breach area and provide accurate regional division data for subsequent disaster analysis, emergency response and terrain modeling.

[0014] Preferably, confirming the segmentation boundary of the breach rough segmentation region map 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 coordinate eigenvalues ​​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.

[0015] The present invention calculates the local feature texture through the formula , which can effectively extract the local texture features of each pixel in the rough segmentation area of ​​the breach. and neighborhood pixels The grayscale value of the rupture segmentation area is used to capture local texture differences and generate texture feature values. This calculation method ensures the accurate extraction of texture features and lays the foundation for the confirmation of segmentation boundaries. 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 simple texture feature pixels and complex texture feature pixels. This distinction method helps to further accurately identify different structures in the breach area, especially the texture features of complex terrain, which helps to better determine the boundaries. 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 segmentation boundaries, avoids the uncertainty of human judgment, and ensures the consistency and accuracy of the segmentation results. Based on the comparison of simple and complex texture feature pixels, the segmentation boundaries of the breach area can be confirmed more carefully, especially in areas with significant texture changes. In this way, the determination of the boundary not only depends on a large range of 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 breach segmentation area boundary data can be generated, providing high-precision boundary information for subsequent breach monitoring, analysis and emergency management. This eigenvalue-based boundary confirmation method effectively improves the reliability of segmentation and is suitable for automated breach analysis in complex environments.

[0016] Preferably, performing layered dynamic image processing on the breach offshore area map and the breach onshore area map comprises 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.

[0017] The present invention can accurately capture the dynamic changes of the breached offshore area, especially the water fluctuation area, through continuous frame acquisition and differential image calculation. By maximizing the inter-class variance calculation, the fluctuation area can be effectively screened out, making the dynamic monitoring of the offshore area more accurate. For the multi-perspective image acquisition and processing of the onshore area, by separating the complex background layer, a clear distinction can be made between the terrain and non-terrain areas, thereby improving the accuracy of image analysis, especially in complex environments. The layered dynamic image processing method can use corresponding algorithms for image processing according to the different characteristics of the offshore and onshore areas, reducing the need for human intervention and improving the degree of automation and intelligence of image processing. By performing layered processing on the offshore and onshore areas, more comprehensive monitoring data of the breach area can be provided, helping analysts to accurately understand the actual situation of the breach area, especially the terrain and dynamic changes, which is helpful for responding to emergencies.

[0018] Preferably, separating the complex background layer of the multi-view image 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.

[0019] The present invention can perform multi-angle analysis on the land area of ​​the breach, especially the spatial height structure annotation, by combining the images of the topographic perspective and the oblique perspective. In this way, the terrain changes in the area can be better determined and the accuracy of the terrain structure can be improved. The terrain and non-terrain layer separation based on the spatial height structure area data can effectively separate the terrain features and complex environment in the background, so that the complex background information such as vegetation and buildings in the non-terrain area will not interfere with the terrain analysis, thereby improving the image resolution. The use of multi-view image processing means, especially the combination of oblique perspective and topographic perspective, helps to capture terrain information more comprehensively, avoid blind spots under a single perspective, and provide a more three-dimensional environmental perception effect. Through automated image center point confirmation and perspective screening, the background layer separation can be quickly performed in a large amount of data, reducing manual intervention, improving image processing efficiency, and accelerating the assessment speed of the disaster area. The terrain and non-terrain separation image finally generated can provide more refined data support for subsequent geographic analysis, disaster assessment and emergency response, and provide a reliable basis for the dynamic changes and feature recognition of the breach area.

[0020] Preferably, step S3 comprises 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: texture fusion is performed on the water breach point cloud image according to the pixel normal vector of the water breach point cloud to generate a water breach texture map.

[0021] The present invention generates a complete dynamic map of the breach area in the water area by splicing the dynamic images of the breach sea area and the land area, and can intuitively display the overall dynamic changes of the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparsely clouding the dynamic map of the breach area in the water area, the two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional spatial analysis. The calculation of the pixel normal vector of the water breach point cloud helps to extract the surface geometric features in the point cloud, especially in a complex breach water environment, and can accurately analyze the geometric information such as the slope and surface changes of the breach area. This provides accurate geometric data support for subsequent structural analysis and modeling. By texture fusion of the pixel normal vector of the water breach point cloud, a high-quality texture image is generated, which can more delicately express 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, and improves the visual performance effect. Through the complete process of image stitching, point clouding, normal vector calculation and texture fusion, the three-dimensional morphological structure of the water breach can be fully and accurately displayed, which provides a scientific basis and data support for disaster assessment, rescue plan formulation and water management.

[0022] In this specification, a system for constructing a real-scene 3D model of a water breach is provided, which is used to execute the above-mentioned method for constructing a real-scene 3D model of a water breach. The system for constructing a real-scene 3D model of a water breach includes: 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 HSV color space conversion on the 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.

[0023] The beneficial effect of the present invention is that by splicing the dynamic images of the breach sea area and the land area, a complete dynamic map of the water breach area is generated, which can intuitively display the overall dynamic changes of the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparsely clouding the dynamic map of the water breach area, the two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional spatial analysis. The calculation of the pixel normal vector of the water breach point cloud helps to extract the surface geometric features in the point cloud, especially in a complex breach water environment, and can accurately analyze the geometric information such as the slope and surface changes of the breach area. This provides accurate geometric data support for subsequent structural analysis and modeling. By texture fusion of the pixel normal vector of the water breach point cloud, a high-quality texture image is generated, which can more delicately express 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, and improves the visual performance effect. Through the complete process of image stitching, point clouding, normal vector calculation and texture fusion, the three-dimensional morphological structure of the water breach can be fully and accurately displayed. This provides a scientific basis and data support for disaster assessment, rescue plan formulation and water 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, dynamic image utilization, and optimized point cloud model construction and texture mapping. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the steps of a method for constructing a real-life three-dimensional model of a water breach; Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart; Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying 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 implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0027] 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 used only to distinguish one unit from another unit. 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 listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 A method for constructing a real-life three-dimensional model of a water breach, the method comprising the following steps: Step S1: Acquire the position data of the breach area control points; perform surveying and mapping benchmark confirmation on the position data of the breach area control points to generate the benchmark data of the breach area control points; perform three-dimensional coordinate conversion on the benchmark data of the breach area control points 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; Step S2: based on the standard UAV satellite signal, drone aerial image acquisition is performed to obtain an aerial image of the breach area; the aerial image of the breach area is converted into an HSV color space to generate a breach area color space conversion image; the standard breach area aerial image is divided into breach sea and land areas using the breach area color space image, thereby generating a breach sea area map and a breach land area map; the breach sea area map and the breach land area map are subjected to layered dynamic image processing to generate a breach sea area dynamic map and a breach land area dynamic map; Step S3: performing 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 breach water area; performing image texture fusion on the dynamic image of the breach water area to generate a texture image of the breach water area; 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.

[0029] The present invention obtains the control point position data of 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 by the surveying and mapping benchmark to ensure the accuracy and reliability of the data and generate the control point benchmark data of the breach area. The benchmark data is used to perform three-dimensional coordinate conversion to generate the plane coordinate data of the breach area. This step is to convert the three-dimensional data into a two-dimensional plane representation to facilitate subsequent processing. The satellite signal of the unmanned aerial vehicle is received based on the plane coordinate data to obtain the standard unmanned aerial vehicle satellite signal to provide navigation support for subsequent image acquisition. This step ensures the accuracy of the control point and lays the foundation for subsequent image acquisition and data processing, thereby improving the accuracy and efficiency of the entire process. Aerial photography is performed using standard unmanned aerial vehicle satellite signals to obtain aerial images of the breach area. The acquired aerial images are converted into HSV color space to better analyze the color information and generate color space conversion images of the breach area. The aerial images are analyzed using color space images to divide the sea area map and the land area map of the breach. The sea area map and the land area map are subjected to layered dynamic image processing to generate dynamic images of the sea area and the land area of ​​the breach. This step achieves accurate identification and classification of the breach area through meticulous color analysis and dynamic processing, providing rich visual information for subsequent image stitching and texture processing. The dynamic images of the breach offshore area and the dynamic images of the land area are stitched to generate a dynamic image of the water breach area. The generated dynamic image is texture fused to obtain a texture map of the water breach to enhance the visual effect. Through image stitching and texture fusion, a high-quality dynamic image of the water breach area is generated, which is convenient for subsequent three-dimensional model construction and analysis, and improves the realism and visualization of the model. The initial point cloud model is constructed based on the point cloud optimization data of the water breach. The initial point cloud model is UV unfolded, and the water breach texture map is mapped to the unfolded model to generate a texture mapping model of the water breach. The texture mapping model is rendered to generate a real-life three-dimensional model of the water breach. This process makes the model not only realistic in shape but also more vivid in visual effect by mapping the texture to 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 precise data acquisition, efficient drone image collection and processing, dynamic image utilization, and optimized point cloud model construction and texture mapping.

[0030] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a method for constructing a real-life 3D model of a water breach according to the present invention. In this example, the method for constructing a real-life 3D model of a water breach comprises the following steps: Step S1: Acquire the position data of the breach area control points; perform surveying and mapping benchmark confirmation on the position data of the breach area control points to generate the benchmark data of the breach area control points; perform three-dimensional coordinate conversion on the benchmark data of the breach area control points 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; In the embodiment of the present invention, multiple control points are arranged in the breach area by using GPS positioning equipment to ensure that the position data of each control point is accurate. The three-dimensional coordinate data (longitude, latitude, altitude) of each control point is recorded for subsequent processing. The acquired control point position data is compared with the existing surveying and mapping benchmark (such as the national surveying and mapping coordinate system). A ground surveying instrument (such as a total station) is used for on-site confirmation to ensure the precise position of the control point, generate the benchmark data of the control point in the breach area, and record the standard coordinates and related information of each control point. Based on the benchmark data of the control point in the breach area, the original three-dimensional coordinates are converted to generate plane coordinate data, and the obtained plane coordinate data of the breach area will be used for subsequent drone operations. The drone is deployed at the confirmed plane coordinate position of the breach area, and the GNSS (Global Navigation Satellite System) module of the drone is ensured to be in good working condition. The drone satellite signal is received to ensure that the drone obtains the standard drone satellite signal and completes the position calibration. The signal strength, number of satellites and their positions received by the drone are recorded to ensure the accuracy of subsequent data collection.

[0031] Step S2: based on the standard UAV satellite signal, drone aerial image acquisition is performed to obtain an aerial image of the breach area; the aerial image of the breach area is converted into an HSV color space to generate a breach area color space conversion image; the standard breach area aerial image is divided into breach sea and land areas using the breach area color space image, thereby generating a breach sea area map and a breach land area map; the breach sea area map and the breach land area map are subjected to layered dynamic image processing to generate a breach sea area dynamic map and a breach land area dynamic map; In an embodiment of the present invention, aerial photography is performed based on the standard drone satellite signal obtained in step S1 by using a drone. Ensure that the drone flies stably in the breach area and perform multi-angle aerial photography according to a preset route. Record the time, location and altitude information of the aerial image for subsequent analysis. Perform HSV (hue, saturation, brightness) color space conversion on the acquired aerial image of the breach area to better analyze the color features in the image. The breach area color space conversion image obtained using a color space conversion formula (such as an RGB conversion formula and an HSV conversion formula) will be used for subsequent area division. Using the color information in the breach area color space image, the standard breach area aerial image is divided into sea and land areas. According to the characteristics of hue (H) and saturation (S), a threshold is set, and the pixels in the image are divided into sea areas and land areas, generating two binary images: a breach sea area map and a breach land area map. Continuous frame image acquisition is performed on the breach sea area map to obtain a continuous frame image of the breach sea area. Multi-view image acquisition is performed on the breach land area map to obtain a multi-view image of the breach land area. The dynamic change analysis of these two types of images is performed separately: the dynamic change of the offshore area is calculated using the image difference method, and the fluctuation area is marked. The complex background is separated from the multi-view images to extract the terrain and non-terrain features.

[0032] Step S3: performing 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 breach water area; performing image texture fusion on the dynamic image of the breach water area to generate a texture image of the breach water area; In an embodiment of the present invention, by ensuring that the dynamic image of the breached sea area and the dynamic image of the breached land area are sorted in time series, their timestamps and perspectives are aligned to achieve seamless splicing. A feature point detection algorithm (such as SIFT or ORB) is used to identify key feature points in each frame of the image. Matching is performed based on the position of the feature points and the descriptor to determine the overlapping area. The transformation matrix of the spliced ​​image is calculated, and the mismatch is eliminated by the RANSAC algorithm to ensure accurate image alignment. The image is geometrically transformed using the transformation matrix to seamlessly splice the dynamic image of the sea area and the dynamic image of the land area. Image blending technology (such as multi-band fusion) is used to process the overlapping area to avoid obvious visible boundaries, ensure a natural transition as a whole, and output the final spliced ​​image to form a dynamic image of the water area breach area. Texture feature extraction is performed on each frame of the spliced ​​dynamic image of the water area breach area. Texture features can be obtained using local binary patterns (LBP), Gabor filters, and other technologies. The extracted texture features are applied to the original image, and the feature performance is enhanced by adjusting the texture mapping parameters (such as intensity and scale). Select a specific image fusion algorithm (such as weighted average, image pyramid fusion, etc.) to fuse the texture features with the original image. This process ensures that key features are retained and details are enhanced. Perform color correction on the fused image to enhance contrast and make the image more natural and detailed. Use Gaussian blur or median filtering to denoise, improve image quality, remove existing artifacts and noise, and output the final water breach texture map, which will be used for further analysis or visualization.

[0033] 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.

[0034] In an embodiment of the present invention, by using the optimized data of the water breach point cloud as a basis, it is ensured that the data is downsampled and denoised 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 breach. Ensure that the model surface is smooth to avoid defects caused by noise or discontinuity, and output the constructed initial point cloud model of the water breach. UV unfold the initial point cloud model to generate a UV coordinate map. UV unfolding can be completed by a specific UV unfolding algorithm (such as cutting method, spherical projection, etc.) to reduce distortion during texture mapping. Ensure that the UV layout is reasonable to avoid overlapping areas and facilitate subsequent texture mapping. Select the previously generated water breach texture map to ensure that the texture quality and resolution are suitable for application to the three-dimensional model. Map the water breach texture map to the initial point cloud model of the water breach after UV unfolding. Use texture coordinates to accurately apply the texture to the model surface. Make texture adjustments to ensure that the texture matches the geometric shape of the model to avoid stretching or compression, and output the water breach texture mapping model that has been texture mapped. Select a suitable rendering engine (such as Blender, Unity, Unreal Engine, etc.) and import the texture mapping model into it. Configure lighting, material properties and other rendering parameters to ensure that the generated real-life model meets the actual effect. Render the model to generate the final real-life 3D model of the water breach. During the rendering process, you can choose different rendering modes (such as real-time rendering, ray tracing, etc.) to improve the visual effect. Post-process the rendering results and apply high dynamic range imaging (HDR) and color correction to improve the visual quality. Save the rendered 3D model to ensure that it can be used for subsequent display and analysis.

[0035] Preferably, step S1 comprises 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.

[0036] In an embodiment of the present invention, multiple control points are arranged in the breach area by selecting appropriate tools (such as GPS receivers and total stations), and their geographic coordinates (longitude, latitude, and elevation) information is recorded. Ensure that the selected control points are representative and evenly distributed for subsequent data processing and analysis. Use the local surveying and mapping benchmark system to benchmark the acquired control point location data to ensure that it meets the standards. Record the confirmation results, including the number, coordinates, accuracy and other information of the benchmark points, and generate the benchmark data of the breach area control points. Use a specific coordinate conversion algorithm (such as geodetic coordinate conversion method and plane coordinate conversion method) to perform three-dimensional coordinate conversion on the control point benchmark data to obtain plane coordinate data (usually expressed in projection coordinates). Ensure that the terrain characteristics and related errors of the region are taken into account during the coordinate conversion process to improve the accuracy of the data. According to the plane coordinate data of the control points, plan the flight path of the drone, including flight altitude, speed, route spacing, etc., to ensure that the entire breach area is covered. Use drone aerial photography software to set up flight missions to ensure the safety and efficiency of drone flights. Install a GNSS receiver on the drone to ensure that it has good satellite signal reception capabilities. Start the drone and start flying. The GNSS receiver receives satellite signals in real time and records the movement trajectory of the drone. During the flight of the drone, the GNSS receiver continuously records the received satellite signals, including timestamp, satellite ID, pseudorange and other information. The collected satellite signals are preprocessed, 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 drone satellite signals to ensure accurate and reliable data.

[0037] Preferably, step S12 comprises 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.

[0038] In the embodiment of the present invention, the longitude, latitude and elevation information of the breach area control point position data are extracted by analyzing the position data of the breach area control point. The reference parameters are set, including the reference ellipsoid, the projection mode, the regional coordinate system, etc., to generate the plane coordinate reference parameter data and the elevation reference parameter data. According to the plane coordinate reference parameter data and the elevation reference parameter data, the regional control point reference framework is constructed, including the selected control point coordinates and their mutual relationship, the breach area control point reference framework data is generated, and all the control point information in the reference framework is recorded to provide a basis for subsequent coordinate conversion and correction. Based on the breach area control point reference framework data, the elevation difference of each control point is calculated. The corresponding correction model (such as terrain correction, meteorological correction, etc.) is applied to obtain the elevation correction value of each control point. The elevation difference correction is applied to the original control point elevation data to generate the breach area control point elevation correction data to ensure the accuracy of the elevation data. According to the three-dimensional zone coordinate conversion rule, a specific mathematical model is used to perform plane coordinate conversion on the control point position data, generate the breach area control point reference data, and record the converted plane coordinate information. On the basis of the generated control point benchmark data, the initial parameters of the three-dimensional coordinates are set, including: Translation: the displacement of each coordinate axis. Rotation: the rotation angle around each coordinate axis. Scaling factor: the scaling ratio of the coordinate data, to obtain the initial parameter data of the three-dimensional coordinates, in preparation for the subsequent seven-parameter solution. According to the initial parameter data of the three-dimensional coordinates, the seven-parameter solution is performed on the position data of the control points in the breach area, including: Three translation solutions: calculate the translation along the X, Y, and Z directions. Three rotation solutions: calculate the rotation angle around the X, Y, and Z axes. Scaling solution: calculate the scaling factor of the coordinate data, generate the seven parameters of the 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 conversion process, and use algorithms to improve the accuracy and reliability of the conversion.

[0039] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: 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: confirm the segmentation boundary of the breach rough segmentation region map to obtain the breach segmentation region boundary data; divide the breach rough segmentation region map into sea and land regions based on the breach segmentation region boundary data, thereby generating the breach sea region map and the 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.

[0040] In an embodiment of the present invention, an aerial photography route is planned based on a standard drone satellite signal to ensure that the drone covers the entire breach area. The drone is started and flies at a fixed altitude, and an aerial image of the breach area is obtained according to the set flight altitude and speed. Ensure that the position and attitude information of the drone is recorded during the aerial photography process for subsequent image registration. The acquired aerial images are subjected to image filtering to remove noise and enhance image quality. Methods such as median filtering or Gaussian filtering can be used to select appropriate parameters to achieve the best effect. Histogram equalization or gamma correction technology is used to improve the brightness and contrast of the image, making the features of the breach area more obvious. Image geometric transformations are performed, including rotation, scaling and cropping, to ensure the spatial consistency of the image and the accuracy of the target area, generate standard breach area aerial images, and provide a clear data basis for subsequent processing. The standard breach area aerial images are converted to HSV color space for better color analysis. The H (hue), S (saturation) and V (brightness) components are extracted for subsequent processing. The hue and saturation information in the breach area color space image is used to perform regional segmentation on the standard breach area aerial images. Apply threshold segmentation or clustering algorithm (such as K-means algorithm) to extract the rough segmentation area related to the breach characteristics, generate the rough segmentation area map of the breach, and clearly identify the target area. Perform boundary detection on the rough segmentation area map of the breach to confirm the outline of the segmentation area. Canny edge detection algorithm or contour extraction technology can be used to obtain clear segmentation area boundary data. Based on the confirmed breach segmentation area boundary data, divide the sea and land areas. Generate the breach sea area map and breach land area map based on the characteristic differences between water bodies and land. Record the division results to ensure the accuracy of the regional division. Perform layered processing on the generated breach sea area map and breach land area map to extract information at different levels. Apply dynamic image processing technology to generate dynamic maps of the breach sea area and breach land area to show changes in different time periods, analyze and visualize the dynamic changes.

[0041] Preferably, using the breach area color space image to segment the breach area of ​​the standard breach area aerial image comprises: 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.

[0042] In an embodiment of the present invention, by processing the color space image of the breach area, the hue, brightness and saturation information of each pixel are extracted to generate three data sets: Pixel hue: represents the color attribute of each pixel, expressed in the form of angle. Pixel brightness: represents the brightness of each pixel, ranging from 0 to 255. Saturation: represents the vividness of the color of each pixel, ranging from 0 to 255. The extracted pixel hue data is statistically analyzed to calculate the hue change amplitude of each pixel. The following formula is used: Hue change amplitude = max(H) − min(H), where HHH represents the hue of the pixel in the breach area, and the hue change amplitude data of the pixel in the breach area is generated to reflect the stability of the color in different areas. According to the hue change amplitude data, the pixel hue is classified. If the hue change amplitude is less than the set threshold, it is marked as a gently changing hue. If the hue change amplitude is greater than or equal to the set threshold, it is marked as a fluctuating hue, and two types of data sets are generated: the gently changing hue of the breach area and the fluctuating hue of the breach area. Using the gently changing hue and saturation data, the standard breach area aerial images are fuzzy segmented for the ocean area. The fuzzy logic method or region growing algorithm is used to extract the ocean area according to the characteristics of hue and saturation to generate a fuzzy segmented image of the ocean area. The threshold method can be used to set the appropriate hue and saturation range to extract the ocean area. According to the fluctuating hue and pixel brightness, the standard breach area aerial images are fuzzy segmented for the land area. Using similar fuzzy logic methods or region growing algorithms, combined with the characteristics of pixel brightness and fluctuating hue, the fuzzy segmented image of the land area is generated. By setting the threshold of brightness and fluctuating changes, the accurate extraction of the land area is ensured. The generated fuzzy segmented images of the ocean area and the fuzzy segmented images of the land area are clustered by pixel color. K-means clustering or other clustering algorithms are used to classify similar pixels according to color characteristics, generate a rough segmented area map of the breach, and identify the characteristics of the breach area and its corresponding color attributes.

[0043] Preferably, confirming the segmentation boundary of the breach rough segmentation region map 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 coordinate eigenvalues ​​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.

[0044] 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: ,in: Represented as pixels The coordinate eigenvalues ​​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; traverse each pixel point in the rough segmentation area map of the breach, extract the values ​​of the central pixel and its neighboring pixels in turn, and calculate the corresponding texture feature value The preset standard texture feature threshold is denoted as For each calculated pixel feature value :if , then mark the pixel as a complex texture feature pixel. If , then the pixel is marked as a simple texture feature pixel. Based on the simple texture feature pixels and complex texture feature pixels, the segmentation boundary of the ulcer rough segmentation area map is further confirmed. Equal to the preset standard texture feature threshold The pixels are marked as segmentation boundary pixels, and the breach segmentation area boundary data is generated, which contains the coordinate information of all pixels marked as segmentation boundary pixels.

[0045] Preferably, performing layered dynamic image processing on the breach offshore area map and the breach onshore area map comprises 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.

[0046] In an embodiment of the present invention, a continuous frame image of the breached offshore area is collected by using an unmanned aerial vehicle or other acquisition equipment to record a series of time-series images to form a continuous frame image sequence of the breached offshore area. A frame is selected from the continuous frame image sequence as a reference frame, which can be the first frame in the time series or a frame under specific conditions. The selected reference frame is named "reference frame image of the breached offshore area". Pixel-level difference calculation is performed on the image of each frame and the reference frame image: ,in Represents the difference image at coordinates The value of Represents the pixel value of the current frame image, Represents the pixel value of the reference frame image. Maximize the inter-class variance calculation to obtain the variance data of the image. Usually, the threshold can be determined using Otsu's method. Calculate the variance data of the breach sea area, denoted as . The variance data of the breach offshore area As a threshold, the differential image of the breach sea area is marked with the fluctuation area image, and the marking rule is set: if , then mark the area as a fluctuation area, generate a dynamic map of the breach offshore area, and display all fluctuation areas. Collect multi-view images of the breach land area from different angles and positions, record images from different perspectives, and form a multi-view image sequence of the breach land area. Process the multi-view 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 breach land area only contains terrain features. The non-terrain separation image of the breach land area contains non-terrain features such as background and objects.

[0047] Preferably, separating the complex background layer of the multi-view image 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.

[0048] In an embodiment of the present invention, an image of a bird's-eye view is extracted from the collected multi-view images. This process can determine whether an image belongs to a bird's-eye view by setting an angle threshold, which is usually set to a near-vertical angle (for example, an angle less than a preset value). The resulting image is called a "bird's-eye view image of the breach land area." The bird's-eye view image of the breach land area is analyzed to identify the center point of the image. The method used can be to calculate the geometric center of the image, using the following formula: ;in and are the horizontal and vertical coordinates of the center point, is the total number of valid pixels in the image, It is a bird's-eye view image. The obtained center point is confirmed as the "center point of the bird's-eye view image of the breach land area". Based on the center point of the bird's-eye view image, the multi-view images of the breach land area are screened for oblique view images. The screening rules are set to select the oblique view images that meet the conditions according to the angle deviation from the bird's-eye view center point. The generated image set is called the "oblique view image of the breach land area". The bird's-eye view image is annotated with the spatial height structure through the oblique view image of the breach land area. Using stereo vision or depth estimation methods, the spatial height information is extracted from the oblique view image to generate spatial height structure area data, which contains the height information of different areas. Based on the spatial height structure area data, the background terrain layer is separated from the multi-view images of the breach land area. The terrain information is separated from the non-terrain information using a hierarchical segmentation algorithm (such as image thresholding, clustering algorithm, etc.). Finally, a separated image of the breach land area is generated, which includes: the terrain separation image of the breach land area focuses on the terrain features. The non-terrain separation image of the breach land area contains background and other object information.

[0049] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes: 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: texture fusion is performed on the water breach point cloud image according to the pixel normal vector of the water breach point cloud to generate a water breach texture map.

[0050] In an embodiment of the present invention, a dynamic image of the breach offshore area and a dynamic image of the breach on land area are spliced. A 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 by perspective transformation (Homography) to generate a complete "dynamic image of the water area breach area". The generated dynamic image of the water area breach area is sparsely point clouded and the image is converted into a point cloud format. Each pixel is converted into a three-dimensional coordinate (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 to generate "water area breach point cloud optimized data", thereby reducing the computational burden of subsequent processing. The normal vector is calculated for each pixel in the water area breach point cloud image. Neighborhood points (such as k neighborhood or fixed radius neighborhood) are used to estimate the normal vector, and the formula is as follows: ;in For point The normal vector of are the coordinates of the neighborhood points, are the coordinates of the initial point, is the number of neighborhood points. Finally, the "water breach point cloud pixel normal vector" is obtained, which provides direction information for each point. Texture fusion is performed on the point cloud image according to the water breach point cloud pixel normal vector. The point cloud is resampled using the normal vector information to map the texture information of the image to 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 "water breach texture map". Finally, ensure that the generated texture map matches the point cloud, providing a clear view for subsequent visualization and analysis.

[0051] In this specification, a system for constructing a real-scene 3D model of a water breach is provided, which is used to execute the above-mentioned method for constructing a real-scene 3D model of a water breach. The system for constructing a real-scene 3D model of a water breach includes: 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 HSV color space conversion on the 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.

[0052] The beneficial effect of the present invention is that by splicing the dynamic images of the breach sea area and the land area, a complete dynamic map of the water breach area is generated, which can intuitively display the overall dynamic changes of the breach area. This provides basic data for the rapid assessment of the disaster area and improves the response efficiency. By sparsely clouding the dynamic map of the water breach area, the two-dimensional image is converted into a three-dimensional point cloud image. Point cloud downsampling further optimizes the quality and density of point cloud data, reduces data redundancy, and improves the efficiency and accuracy of three-dimensional spatial analysis. The calculation of the pixel normal vector of the water breach point cloud helps to extract the surface geometric features in the point cloud, especially in a complex breach water environment, and can accurately analyze the geometric information such as the slope and surface changes of the breach area. This provides accurate geometric data support for subsequent structural analysis and modeling. By texture fusion of the pixel normal vector of the water breach point cloud to generate a high-quality texture image, the surface features of the water area and the breach area can be more delicately expressed, providing rich texture details for subsequent three-dimensional visualization and rendering, and improving the visual performance effect. Through the complete process of image stitching, point clouding, normal vector calculation and texture fusion, the three-dimensional morphological structure of the water breach can be fully and accurately displayed. This provides a scientific basis and data support for disaster assessment, rescue plan formulation and water 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, dynamic image utilization, and optimized point cloud model construction and texture mapping.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented 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; the aerial image of the breach area is converted into an HSV color space to generate a breach area color space conversion image; the standard breach area aerial image is divided into breach sea and land areas using the breach area color space image, thereby generating a breach sea area map and a breach land area map; the breach sea area map and the breach land area map are subjected to layered dynamic image processing to generate a breach sea area dynamic map and a breach land area dynamic map; Step S3: performing 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 breach water area; performing image texture fusion on the dynamic image of the breach water area to generate a texture image of the breach water area; 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 coordinate eigenvalues ​​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 6, 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. The method for constructing a real-scene three-dimensional model of a water breach according to claim 1, characterized in that: 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: texture fusion is performed on the water breach point cloud image according to the pixel normal vector of the water breach point cloud to generate a water breach texture map.

10. 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 HSV color space conversion on the 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