A dynamic terrain monitoring method and system based on image recognition

Abstract

The present invention relates to the field of geographic image recognition technology, and in particular to a dynamic terrain monitoring method and system based on image recognition. The method comprises the following steps: obtaining survey area data; confirming the geographic coordinates of the breach in the survey area data to obtain the geographic coordinates of the breach; collecting daytime drone oblique photography images of the breach geographic coordinates to generate daytime breach photography images; encrypting the daytime breach photography images in three-dimensional space to generate daytime breach three-dimensional space encrypted data; constructing a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-life three-dimensional model; collecting nighttime drone point cloud data based on the breach geographic coordinates data to generate nighttime breach laser point cloud data. The present invention improves the accuracy and comprehensiveness of dynamic terrain monitoring through precise geographic coordinate confirmation, efficient day and night data collection, multi-dimensional model construction, dynamic trend analysis and visual result display.

Description

Technical Field

[0001] The present invention relates to the field of geographic image recognition technology, and in particular to a dynamic terrain monitoring method and system based on image recognition. Background Art

[0002] Initially, terrain monitoring mainly relied on manual measurement and traditional photogrammetry techniques. However, this method is inefficient and difficult to obtain data in real time. With the development of remote sensing technology, satellite images and aerial photography have begun to be widely used, enabling terrain monitoring to cover larger areas and increase the frequency of data acquisition. With the rapid advancement of information technology and image processing technology, dynamic terrain monitoring methods based on image recognition have gradually emerged. The application of computer vision has made image analysis and feature extraction possible, especially the introduction of deep learning algorithms, which has greatly improved the accuracy and efficiency of image recognition. For example, the successful application of convolutional neural networks (CNNs) in image classification and target detection has provided a new solution for dynamic terrain monitoring. In recent years, the technology combining unmanned aerial vehicles (UAVs) and sensor networks has further promoted the development of dynamic terrain monitoring. UAVs can obtain high-resolution terrain images in a short time and analyze terrain changes through image recognition technology. However, the current traditional terrain monitoring methods cannot effectively collect data under different lighting conditions. At the same time, the lack of a unified monitoring model for water and land makes the monitoring results more three-dimensional and comprehensive, resulting in low accuracy and comprehensiveness of terrain monitoring. Summary of the invention

[0003] Based on this, it is necessary to provide a dynamic terrain monitoring method and system based on image recognition to solve at least one of the above technical problems.

[0004] To achieve the above object, a dynamic terrain monitoring method based on image recognition is provided, the method comprising the following steps:

[0005] Step S1: Acquire survey area data; confirm the geographic coordinates of the breach on the survey area data to obtain the geographic coordinates of the breach; collect daytime drone oblique photography images on the breach geographic coordinates data to generate daytime breach photography images; perform three-dimensional spatial encryption on the daytime breach photography images to generate daytime breach three-dimensional spatial encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional spatial encrypted data to generate a daytime breach real-scene three-dimensional model;

[0006] Step S2: collecting nighttime drone point cloud data according to the breach geographic coordinate data to generate nighttime breach laser point cloud data; measuring the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; constructing a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; performing water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitoring the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data to generate breach land terrain monitoring data;

[0007] Step S3: Acquire breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a water-land integrated digital elevation model of the breach;

[0008] Step S4: Project the three-dimensional coordinates of the feature line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring of the breach.

[0009] The present invention obtains data from the survey area to ensure that the basic information of data collection is accurate, laying a good foundation for subsequent steps. By confirming the geographic coordinates of the breach, the positioning accuracy of the monitoring data is improved and the geographic deviation is reduced. Using drones for oblique photography can efficiently obtain high-resolution terrain images covering a wider area. Encrypting the image in three-dimensional space helps to protect the security of the data and avoid information leakage. The generated real-life three-dimensional model provides an intuitive visualization of the survey area, which helps to better understand the terrain features. Combined with night laser point cloud technology, reliable terrain data can be obtained under different lighting conditions, improving the integrity of the data. By measuring the real-life three-dimensional model, accurate mouth width data is obtained, which provides necessary parameters for subsequent monitoring, and generates a nighttime water digital surface model, which helps to analyze the changes in the water area and its impact on the surrounding environment. By analyzing the water surface elevation trend data, the water level changes are monitored in real time, providing a scientific basis for water area management and decision-making. Combined with the mouth width and water surface height trend data, comprehensive monitoring of the breach is achieved to ensure the accuracy and timeliness of the monitoring data. Flow rate data provides a basis for hydrological analysis and helps to understand the impact of water flow on terrain. By combining flow rate and geographic coordinate data, underwater terrain can be accurately measured to supplement the lack of data above the water surface. By correcting the water level of the measured data, the accuracy of the underwater terrain data is ensured and the reliability of the model is improved. Underwater terrain monitoring realizes the monitoring of underwater terrain and provides important data support for ecological protection and water area management. Combining land and underwater data, an integrated digital elevation model is generated, which provides a global perspective for comprehensive analysis. The generated three-dimensional line drawing can intuitively display the terrain features and help identify important terrain changes. Through visual design, the monitoring results are more readable and easy for non-professionals to understand. The three-dimensional digital line drawing finally generated provides a practical tool for dynamic monitoring, which can support timely response and decision-making and enhance the effectiveness of monitoring. Therefore, the present invention improves the accuracy and comprehensiveness of dynamic terrain monitoring through accurate geographic coordinate confirmation, efficient day and night data collection, multi-dimensional model construction, dynamic trend analysis and visual result display.

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

[0011] Step S11: Acquire survey area data;

[0012] Step S12: confirming the geographical coordinates of the breach on the survey area data to obtain the geographical coordinate data of the breach; collecting daytime drone oblique photography images based on the geographical coordinate data of the breach to generate daytime breach photography images;

[0013] Step S13: performing image preprocessing on the daytime ulcer photography image to generate a standard daytime ulcer photography image, wherein the image preprocessing includes image filtering, image brightness enhancement and image resolution enhancement;

[0014] Step S14: encrypting the standard daytime breach photography image in three-dimensional space to generate daytime breach three-dimensional space encrypted data; constructing a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model.

[0015] The present invention can not only locate the breach position by confirming the geographical coordinates of the breach, but also ensure that the subsequent UAV data collection matches the actual geographic situation, avoiding the deviation problem in traditional manual surveying and mapping. Multi-angle oblique photography based on UAV can cover various terrains and provide a high-precision data source for multi-angle three-dimensional reconstruction. Using UAV for daytime oblique photography (step S12) can not only quickly obtain a large range of high-resolution images, but also reduce the time cost and manpower investment of traditional ground measurement. UAVs can automatically plan flight paths, reduce manual intervention, and improve work efficiency, which is particularly suitable for emergency response in disaster scenes such as breaches. Image preprocessing (step S13) not only improves the visual clarity of the image through image filtering, brightness enhancement and resolution enhancement technologies, but also effectively removes noise and lighting differences during the shooting process, ensuring that the image can maintain high quality under various weather or lighting conditions. In addition, these preprocessings also provide high-quality input data for the subsequent three-dimensional modeling process, reducing the risk of error propagation. Three-dimensional space encryption is not only used to prevent unauthorized data access, but also to ensure data integrity during transmission and storage. For data involving sensitive geographic areas or disaster-sensitive data, this encryption step effectively protects data privacy and security while meeting regulatory requirements. By constructing a real-life 3D model of the breach during the day (step S14), the process can generate a high-precision 3D model that accurately restores geographic and physical features. The 3D model is not only used for visual presentation, but also can be used in engineering calculations, risk assessments, and resource planning when combined with geographic coordinates and spatially encrypted data. Using the 3D model of the breach, the possibility of hydrological changes, soil subsidence, or collapse in the damaged area can be more accurately assessed. Through 3D spatial encryption (step S14), breach data and models can be shared more securely between different departments or institutions, promoting cross-departmental collaboration.

[0016] Preferably, step S14 comprises the following steps:

[0017] Step S141: performing image segmentation on the standard daytime ulcer photography image to generate a ulcer region segmentation image; performing ulcer feature extraction on the ulcer region segmentation image to obtain daytime ulcer region feature data; converting the daytime ulcer region feature data into data point cloud to generate daytime ulcer region feature point cloud data;

[0018] Step S142: performing point cloud density analysis on the daytime breach area feature point cloud data to generate regional feature point cloud density data; performing data block segmentation on the daytime breach area feature point cloud data according to the regional feature point cloud density data to generate regional feature point cloud segmentation data blocks;

[0019] Step S143: encrypting the regional feature point cloud segmentation data blocks to generate regional feature point cloud encrypted data blocks; performing three-dimensional spatial arrangement confusion on the regional feature point cloud encrypted data blocks to generate three-dimensional spatial encrypted data of the daytime breach; performing three-dimensional gridding on the three-dimensional spatial encrypted data of the daytime breach to generate a preliminary three-dimensional grid model of the daytime breach;

[0020] Step S144: extract color and texture information from the standard daytime breach photography image to obtain daytime breach color data and daytime breach texture information data; perform texture color mapping on the preliminary three-dimensional mesh model of the daytime breach using the daytime breach color data and daytime breach texture information data to generate a real-scene three-dimensional model of the daytime breach.

[0021] The present invention can accurately identify the boundaries and internal features of the breach area through image segmentation and feature extraction in step S141. This refined regional segmentation ensures that the data processed subsequently is more accurate, avoids interference with irrelevant areas, and thus enhances the authenticity of the three-dimensional model. The feature point clouding of the breach area (step S141) can convert the two-dimensional image into three-dimensional spatial data. The point cloud data accurately reflects the geometric shape of the breach, especially in complex terrain conditions, and the point clouding technology provides a basis for the accuracy of the model. Through the point cloud data, the model can analyze the shape and changes of the breach from multiple angles. Step S142 can optimize the structure of the point cloud data through point cloud density analysis. Density analysis helps to identify the parts that need to be focused on in the breach area and segment the data to make data processing more efficient. The segmented data blocks help to reduce data redundancy and improve subsequent encryption and processing speed. Encrypting the regional feature point cloud segmentation data blocks (step S143) greatly enhances the security of the data, especially during data transmission or storage. The three-dimensional spatial arrangement confusion further improves the encryption effect and prevents unauthorized access or tampering. The three-dimensional meshing step in step S143 generates a preliminary three-dimensional mesh model based on the encrypted point cloud data. This three-dimensional meshing can not only quickly construct the spatial structure of the breach area, but also provide a solid foundation for subsequent texture and color mapping, ensuring the accuracy and fidelity of the model. Step S144 further enriches the visual effect of the three-dimensional model by extracting color and texture. The mapping of the daytime breach color data and texture information data makes the final model not only have accurate geometric shapes, but also can truly restore the visual characteristics of the breach area.

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

[0023] Step S21: using a laser radar drone to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data;

[0024] Step S22: measuring the opening width of the daytime breach real-scene three-dimensional model to obtain the opening width data of the daytime breach; confirming the fuzzy area of ​​the nighttime breach on the laser point cloud data of the nighttime breach according to the opening width data of the daytime breach to obtain the point cloud data of the fuzzy area of ​​the nighttime breach; constructing a digital model of the surface of the nighttime water area based on the point cloud data of the fuzzy area of ​​the nighttime breach to generate a digital surface model of the nighttime water area;

[0025] Step S23: performing night breach water surface height analysis on the night water area digital surface model to generate night breach water surface height data; performing water surface elevation raster analysis on the night breach water surface height data to generate night breach water surface height trend data;

[0026] Step S24: Monitor the land topography of the breach through the daytime breach gate width data and the nighttime breach water surface height trend data, thereby generating the land topography monitoring data of the breach.

[0027] The present invention collects night point cloud data of the breach by using a laser radar drone, and can ensure the acquisition of high-precision point cloud data even under insufficient light conditions. This method effectively makes up for the limitations of daytime data, so that breach monitoring and analysis can be carried out under all-weather conditions. Step S22 combines the daytime breach gate width data with the nighttime laser point cloud data, which helps to identify the fuzzy area of ​​the breach at night. In this way, not only can the night breach area be confirmed more accurately, but also a solid data foundation can be provided for the construction of the digital model of the night water area, so as to achieve seamless integration of day and night data and improve the comprehensiveness of the data. The digital model of the night water surface (step S22) can accurately reflect the geometric shape of the water surface. This is particularly important for monitoring changes in the water area near the breach, especially in floods or other hydrological events, where changes in the water surface indicate potential dangers. The water surface model constructed by point cloud data provides accurate basic data for subsequent analysis and decision-making. Step S23 performs water surface height analysis on the digital surface model of the night water area, providing accurate water surface height information. This not only helps to monitor water level changes in real time, but also generates trend data through elevation raster analysis, which helps predict future changes in water surface height. This trend data is crucial for risk assessment of the breach area and helps prevent further damage caused by rising water levels. Step S24 provides dynamic monitoring of the breach land terrain by combining the daytime gate width data with the nighttime water surface height trend data. This monitoring can effectively reflect the interaction between water and land, and help better understand the terrain changes in the breach area. For example, when the water level rises, erosion or subsidence of the land can be discovered more quickly, so that corresponding measures can be taken in time. Through the generation and analysis of nighttime water surface height trend data, the future hydrological conditions in the breach area can be effectively predicted. This predictive capability is not only helpful for short-term decision-making, but also can provide data support for long-term breach repair and planning.

[0028] Preferably, step S23 includes the following steps:

[0029] Step S231: performing rough water surface area extraction on the nighttime water area digital surface model based on a preset elevation range to obtain rough water surface area data;

[0030] Step S232: performing regional reflectivity difference calculation on the rough water surface area data to obtain regional reflectivity difference data; performing regional boundary adjustment on the rough water surface area data by using the regional reflectivity difference data, thereby generating night breach water surface area data;

[0031] Step S233: selecting water surface key points from the night breach water surface area data to obtain the night breach water surface area key point data, wherein the water surface key points selected include the breach center area, the breach left area and the breach right area; performing water surface elevation profile height comparison on the night breach water surface area data according to the night breach water surface area key point data to generate night breach water surface height data; performing grid unit conversion on the night breach water surface height data to generate water surface height grid units;

[0032] Step S234: Calculate the local elevation slope of the water surface height grid unit to obtain the local elevation slope data of the water surface; perform curve conversion on the local elevation slope data of the water surface to generate a water surface elevation slope curve; perform trend analysis on the water surface elevation slope curve to generate nighttime breach water surface height trend data.

[0033] The present invention can quickly identify the water surface area by roughly extracting the digital surface model of the night water area through a preset elevation range. Although it is rough water surface area data, this step provides a basis for subsequent fine boundary adjustment and significantly improves processing efficiency. The boundary of the rough water surface area data is adjusted by using the regional reflectivity difference, which improves the accuracy of the water surface area. The reflectivity difference calculation is based on the characteristics of water body reflecting light, so that the system can accurately identify the boundary between the water surface and the land. This method reduces the traditional manual correction error and ensures that the generated night breach water surface area data is more reliable. Step S233 selects the water surface key point data of the breach center, left side and right side, and by comparing the elevation profiles of these key points, the change of the breach water surface height can be accurately reflected. This elevation profile comparison method captures the slight fluctuations of the water surface in space, helping to better understand the detailed changes in the water surface elevation. Step S233 converts the water surface height data into grid units, and in this way, the data can be more easily calculated and converted into curves in the subsequent slope. Rasterized data makes the processing of elevation data more standardized and modular, which not only improves the efficiency of data storage and analysis, but also facilitates the fusion and comparison of results with other rasterized geographic information data. Step S234 calculates the local elevation slope of the water surface height grid unit, which can reveal the change of the water surface slope in the breach area. The local elevation slope data can reflect the inclination of the breach water surface in the local area, providing an important basis for understanding the flow trend of water bodies or terrain changes. The local elevation slope data generates a water surface elevation slope curve through curve conversion, which further improves the visualization of the data. This curve shows the slope characteristics of the water surface height changing with space, which helps to discover subtle fluctuations in the water surface elevation, especially under dynamic changes such as floods and tides, and can more clearly capture the fluctuation of the water surface.

[0034] Preferably, step S232 includes the following steps:

[0035] Step S2321: performing Sobel operator calculation on the rough water surface area data to generate local reflectivity gradient data of the rough water surface area; marking high gradient areas on the local reflectivity gradient data of the rough water surface area to generate high gradient area data of rough water surface reflectivity;

[0036] Step S2322: constructing a gradient enhancement map for the local reflectivity gradient data of the rough water surface area according to the rough water surface reflectivity high gradient area data, generating a water surface area reflectivity gradient enhancement map; performing adjacent gradient difference calculation on the water surface area reflectivity gradient enhancement map, obtaining regional reflectivity difference data;

[0037] Step S2323: performing gradient peak tracking on the rough water surface area data through the regional reflectivity difference data to generate regional gradient peak data; performing boundary smoothing on the rough water surface area data through the regional gradient peak data to generate boundary adjustment data for the breach water surface area at night;

[0038] Step S2324: Perform regional consistency verification on the rough water surface area data based on the nighttime breach water surface area boundary adjustment data, thereby generating nighttime breach water surface area data.

[0039] The present invention generates local reflectivity gradient data through Sobel operator calculation and identifies high gradient areas. This step can accurately identify the boundary of the water surface area, especially the reflectivity difference between the water surface and the surrounding objects. Through gradient analysis, the system can effectively avoid environmental noise interference and improve the accuracy of the boundary of the water surface area. In step S2322, based on the high gradient area data, a reflectivity gradient enhancement map is constructed to amplify the difference in water surface reflectivity gradient. Gradient enhancement processing helps to further refine the identification of the boundary, making the boundary of the water surface area clearer and ensuring that the subsequent adjacent gradient difference calculation is more accurate. The adjacent gradient difference calculation of step S2322 enables the system to identify subtle changes in reflectivity in the water surface area. Through this process, water surface features with different reflectivities can be better distinguished, the resolution of the data is enhanced, and the regional reflectivity difference data is further optimized. This step lays a solid foundation for subsequent boundary adjustment. Step S2323 can determine the precise boundary of the water surface area by tracking the regional gradient peak. Gradient peak tracking technology can ensure the smoothness of the boundary and avoid excessive tortuosity or irregularity of the boundary line, thereby generating more continuous and natural boundary data. Through this boundary optimization, the system greatly improves the recognition quality of the water surface area. After gradient peak tracking, the system performs boundary smoothing on the rough water surface area data, so that the generated night breach water surface area boundary is more consistent with the actual terrain conditions. The smoothing process not only improves the stability of the boundary line, but also reduces the error in subsequent model construction, which helps to generate a more stable water surface digital model. Step S2324 ensures the reliability of the final generated night breach water surface area data through consistency verification. The regional consistency verification process can identify and correct existing local errors to ensure that the water surface area data is consistent at different resolutions. This process ensures the comprehensiveness and accuracy of the generated data and reduces inconsistencies or anomalies in the data.

[0040] Preferably, step S3 comprises the following steps:

[0041] Step S31: obtaining breach flow velocity data;

[0042] Step S32: laying out the underwater topography survey plan line based on the breach flow velocity data and the breach geographic coordinate data, and generating an underwater topography survey plan line layout map; performing underwater topography survey of the breach on the underwater topography survey plan line layout map using the cross-section method based on the depth-measuring unmanned ship, and generating the breach underwater topography survey data;

[0043] Step S33: perform water depth verification on the breach underwater topography measurement data to generate breach underwater topography verification data; perform water level correction on the breach underwater topography verification data to generate complete underwater topography data of the breach area;

[0044] Step S34: Perform underwater terrain monitoring on the complete underwater terrain data of the breach area to generate underwater terrain monitoring data of the breach; construct a digital elevation model using the breach land terrain monitoring data and the breach underwater terrain monitoring data to generate a land and water integrated digital elevation model of the breach.

[0045] The present invention successfully constructs a complete integrated water-land digital elevation model by combining the breach flow velocity data, underwater topography measurement data and land topography monitoring data through steps S31 to S34. This model provides comprehensive three-dimensional topographic information for the breach area, can accurately reflect the changing trend of underwater and land topography, and improves the surveying and mapping accuracy. In step S32, the underwater topography measurement plan line is laid out using the breach flow velocity data and geographic coordinate data, ensuring the rationality and efficiency of the measurement plan. By scientifically laying out the measurement plan line, the key topographic features of the breach area can be covered, making the subsequent depth measurement operation more accurate, and improving the comprehensiveness and reliability of the underwater topography measurement. The cross-section method is used to measure the underwater topography of the laid survey plan line, and the generated breach underwater topography measurement data provides higher accuracy in the vertical direction. This method ensures the accuracy of underwater topography measurement by seamlessly connecting water and land topography, and is particularly suitable for measurement in complex terrain areas of the breach. By performing water depth verification and water level correction on the breach underwater topography measurement data, more accurate breach underwater topography verification data is generated. This correction process effectively eliminates measurement deviations caused by water level fluctuations or equipment errors, ensuring that the final generated underwater terrain data is more reliable and consistent with the actual underwater conditions. The complete underwater terrain data of the breach area is combined with the land terrain monitoring data to achieve multi-level monitoring of water and land terrain. The multi-dimensional terrain data underwater and on land provide the system with comprehensive terrain information of the breach area, ensuring the integrity and consistency of the data during the construction of the digital elevation model. The integration of underwater and onshore terrain monitoring data of the breach can support real-time tracking of terrain changes in complex environments. Through these monitoring data, the system can identify dynamic changes in the terrain of the breach area, and provide scientific references for disaster prevention and control, engineering design and other applications in a timely manner, which has broad application prospects.

[0046] Preferably, constructing a digital elevation model using breach land terrain monitoring data and breach underwater terrain monitoring data includes:

[0047] Perform point cloud conversion on the breach land terrain monitoring data to generate breach land terrain monitoring point cloud data; perform ground attachment filtering on the breach land terrain monitoring point cloud data to generate breach land terrain monitoring filtered data, wherein the ground attachment filtering includes vegetation, water surface and vehicles;

[0048] Extract spatial three-dimensional features from the breach underwater terrain monitoring data and breach land terrain monitoring filtered data to obtain breach land terrain spatial feature data and breach underwater terrain spatial feature data; select spatial common points from the breach land terrain spatial feature data and breach underwater terrain spatial feature data to obtain land and underwater spatial boundary point data;

[0049] The vertex elevations of the onshore and underwater spatial boundary point data are triangulated to generate onshore and underwater triangular meshes; polygonal boundary overlap adjustment is performed on the onshore and underwater triangular meshes to generate onshore and underwater triangular mesh adjustment data; spatial interpolation is performed on the onshore and underwater triangular meshes based on the onshore and underwater triangular mesh adjustment data to generate a land and water integrated digital elevation model of the breach.

[0050] The present invention generates a high-precision land-water integrated digital elevation model by extracting spatial features and constructing triangular meshes for land and water terrain monitoring data. The model can accurately reflect the terrain characteristics of the breach area, including the height changes of complex underwater terrain and land terrain, and improves the comprehensiveness and accuracy of terrain modeling. The land terrain monitoring data is filtered for ground attachments, and attachments such as vegetation, water surface and vehicles that affect the accuracy of the terrain are effectively eliminated. This step ensures that the monitoring data only reflects the actual characteristics of the terrain, improves the purity of the data, and thus reduces the errors in modeling. Through the selection of spatial common points, the seamless connection of land and underwater terrain data is achieved. The land and underwater junction point data provides an accurate boundary basis for the land-water integrated model, ensures the natural transition of the land-water connection area, and eliminates the break problem between different terrain data. The land and underwater terrain junction points are triangulated and the boundary overlap adjustment is performed to generate a refined triangular mesh structure. This detailed triangulation helps to more accurately capture complex terrain changes during elevation modeling, especially the height differences of slopes, potholes and bottoms in the breach area. The boundary discontinuity problem between the onshore and underwater terrain grids was solved by overlapping polygon boundaries, further improving the overall consistency of the model. This ensures smooth connection between onshore and underwater triangular grids and eliminates the sudden change or incoordination caused by terrain data fusion. The application of spatial interpolation technology further refines the integrated land-water digital elevation model. Through the interpolation processing of onshore and underwater triangular grids, a smoother and more complete elevation model is generated, which is particularly suitable for the dynamic changes of complex terrain in the breach area and enhances the application ability of the model in different scenarios. The integrated land-water digital elevation model provides reliable data support for application scenarios such as disaster warning, engineering planning and flood control in the breach area. Accurate land-water terrain information helps to analyze the water flow diffusion, erosion risk and structural stability in the breach area, thus providing a scientific basis for engineering design and emergency response. Through the meticulous processing of triangular grids and spatial interpolation technology, the deviation problem of onshore and underwater terrain data at the boundary transition is solved, ensuring the high consistency and accuracy of the model. This reduces the errors caused by inconsistent data and provides a stable basis for further analysis.

[0051] Preferably, step S4 comprises the following steps:

[0052] Step S41: extracting contour lines from the breach water-land integrated digital elevation model to obtain the breach terrain contour lines; extracting water systems and shorelines from the breach waterway integrated digital elevation model to obtain water body boundary lines and breach shoreline layers;

[0053] Step S42: extracting important terrain feature lines from the breach water-land integrated digital elevation model to generate a breach important terrain feature line set, wherein the important terrain feature lines extracted include breach edge lines, steep slope area boundary lines, depression lines, and ridge lines;

[0054] Step S43: Perform three-dimensional coordinate projection on the breach terrain contour lines, water body boundary lines, breach shoreline layers, and important terrain feature line sets of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring operations on the breach.

[0055] The present invention extracts contour lines from the integrated water-land digital elevation model to clearly display the terrain changes in the breach area, including the distribution of terrain at different heights. This provides an accurate terrain contour basis for dynamic terrain monitoring of the breach, and is convenient for analyzing the terrain trend of the water-land boundary area. By extracting the water system and the coastline, the water body boundary line and the breach coastline layer are generated, and the water body and land boundaries in the breach area are accurately divided. This provides basic data support for subsequent water flow diffusion simulation and hydrological analysis, and is also helpful for disaster prevention and control and breach repair planning. The extracted breach edge line, steep slope area boundary line, depression line and ridge line and other important terrain feature line sets can intuitively display the key geomorphic features around the breach. In particular, the extraction of the steep slope area boundary line is helpful to assess the landslide or collapse risk in the breach area. The contour lines, water body boundary lines, coastlines and important terrain feature lines in the breach area are projected into three-dimensional coordinates to generate a three-dimensional line drawing of the breach, which provides an intuitive three-dimensional representation for spatial monitoring and analysis of the breach. This three-dimensional image can help engineers better understand the topographic structure of the breach, optimize breach protection design and water conservancy project planning, and the generated three-dimensional digital line drawing of the breach has dynamic terrain monitoring capabilities. By setting the color and style of the lines, it is easier to monitor the terrain changes in the breach area at different time periods. The diversified settings of line drawing styles, such as marking slope changes with different colors, can significantly improve the intuitiveness and analysis efficiency of breach area monitoring. The three-dimensional digital line drawing can not only intuitively display the terrain changes in the breach area, but also combine elements such as water body boundaries and contour lines for further multi-dimensional analysis, such as the impact of rising water levels on the breach area and the prediction of flood diffusion paths. This multi-dimensional data integration helps to provide more comprehensive analysis support.

[0056] In this specification, a dynamic terrain monitoring system based on image recognition is provided, which is used to execute the above-mentioned dynamic terrain monitoring method based on image recognition. The dynamic terrain monitoring system based on image recognition includes:

[0057] The daytime land measurement module is used to obtain survey area data; confirm the breach geographic coordinates of the survey area data to obtain the breach geographic coordinates data; collect the breach geographic coordinates data by daytime drone oblique photography to generate daytime breach photography images; encrypt the daytime breach photography images in three-dimensional space to generate daytime breach three-dimensional space encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model;

[0058] The nighttime land measurement module is used to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data; measure the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; construct a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; perform water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitor the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data, thereby generating breach land terrain monitoring data;

[0059] The land-water integrated monitoring module is used to obtain the breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a land-water integrated digital elevation model of the breach;

[0060] The dynamic terrain monitoring module is used to perform three-dimensional coordinate projection of the characteristic line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line drawing of the breach; set the line color and style of the three-dimensional line drawing of the breach to generate a three-dimensional digital line drawing of the breach to perform dynamic terrain monitoring operations of the breach.

[0061] The beneficial effect of the present invention is that by confirming the geographic coordinates of the breach in the survey area data, the data obtained is ensured to be highly accurate, laying the foundation for subsequent monitoring and analysis. The daytime breach photography images generated by using drones for daytime oblique photography can provide rich surface information and support the subsequent three-dimensional model construction. The three-dimensional model is constructed based on the three-dimensional spatial encrypted data of the daytime breach, and the generated daytime breach real-life three-dimensional model can intuitively display the terrain characteristics of the breach area, providing a basis for further analysis. According to the breach geographic coordinate data, nighttime drone point cloud data is collected, and the generated nighttime breach laser point cloud data can capture detailed information about the nighttime water area, supplementing the shortcomings of daytime collection. The mouth width of the daytime real-life three-dimensional model is measured, and the generated mouth width data helps to evaluate the degree of opening of the breach and the dynamics of water flow. The digital surface model of the nighttime water area is generated based on the nighttime laser point cloud data, and the water surface state and its changes can be analyzed. Through the water surface elevation raster analysis, the generated nighttime breach water surface height trend data provides necessary data support for hydrological analysis and flow rate evaluation. The breach land terrain monitoring data generated by integrating the daytime breach gate width data and nighttime water surface height trend data can effectively reflect the terrain changes in the breach area. Obtaining the breach flow velocity data provides necessary parameters for underwater terrain measurement, which helps to accurately evaluate the flow characteristics. Underwater terrain measurement is performed based on the breach flow velocity data and geographic coordinates, and the generated measurement data provides a basis for subsequent water level correction and monitoring. After water level correction, the generated complete underwater terrain data provides high-quality data support for underwater feature analysis. Combining the onshore and underwater terrain monitoring data, the generated digital elevation model can fully reflect the water-land boundary characteristics of the breach area. The three-dimensional coordinate projection of the feature line set of the breach water-land integrated digital elevation model is performed, and the generated three-dimensional line drawing can intuitively display the terrain characteristics and is easy to understand. By setting the line color and style, the generated three-dimensional digital line drawing of the breach is more visually intuitive, which is helpful for performing dynamic terrain monitoring operations of the breach. This systematic monitoring process provides decision makers with efficient information support through detailed data collection and analysis, which helps to quickly respond to potential terrain changes or disaster risks. The generated three-dimensional digital line map can not only be used for water conservancy project monitoring, but also provide support for geological disaster assessment, environmental protection and infrastructure construction, and has broad application prospects. Therefore, the present invention improves the accuracy and comprehensiveness of dynamic terrain monitoring through precise geographic coordinate confirmation, efficient day and night data collection, multi-dimensional model construction, dynamic trend analysis and visual result display. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A schematic diagram of the steps of a dynamic terrain monitoring method based on image recognition;

[0063] Figure 2 for Figure 1Detailed implementation steps of step S2 in the flowchart;

[0064] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0065] 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

[0066] 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.

[0067] 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.

[0068] 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.

[0069] To achieve this, please refer to Figures 1 to 3 , a dynamic terrain monitoring method based on image recognition, the method comprising the following steps:

[0070] Step S1: Acquire survey area data; confirm the geographic coordinates of the breach on the survey area data to obtain the geographic coordinates of the breach; collect daytime drone oblique photography images on the breach geographic coordinates data to generate daytime breach photography images; perform three-dimensional spatial encryption on the daytime breach photography images to generate daytime breach three-dimensional spatial encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional spatial encrypted data to generate a daytime breach real-scene three-dimensional model;

[0071] Step S2: collecting nighttime drone point cloud data according to the breach geographic coordinate data to generate nighttime breach laser point cloud data; measuring the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; constructing a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; performing water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitoring the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data to generate breach land terrain monitoring data;

[0072] Step S3: Acquire breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a water-land integrated digital elevation model of the breach;

[0073] Step S4: Project the three-dimensional coordinates of the feature line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring of the breach.

[0074] The present invention obtains data from the survey area to ensure that the basic information of data collection is accurate, laying a good foundation for subsequent steps. By confirming the geographic coordinates of the breach, the positioning accuracy of the monitoring data is improved and the geographic deviation is reduced. Using drones for oblique photography can efficiently obtain high-resolution terrain images covering a wider area. Encrypting the image in three-dimensional space helps to protect the security of the data and avoid information leakage. The generated real-life three-dimensional model provides an intuitive visualization of the survey area, which helps to better understand the terrain features. Combined with night laser point cloud technology, reliable terrain data can be obtained under different lighting conditions, improving the integrity of the data. By measuring the real-life three-dimensional model, accurate mouth width data is obtained, which provides necessary parameters for subsequent monitoring, and generates a nighttime water digital surface model, which helps to analyze the changes in the water area and its impact on the surrounding environment. By analyzing the water surface elevation trend data, the water level changes are monitored in real time, providing a scientific basis for water area management and decision-making. Combined with the mouth width and water surface height trend data, comprehensive monitoring of the breach is achieved to ensure the accuracy and timeliness of the monitoring data. Flow rate data provides a basis for hydrological analysis and helps to understand the impact of water flow on terrain. By combining flow rate and geographic coordinate data, underwater terrain can be accurately measured to supplement the lack of data above the water surface. By correcting the water level of the measured data, the accuracy of the underwater terrain data is ensured and the reliability of the model is improved. Underwater terrain monitoring realizes the monitoring of underwater terrain and provides important data support for ecological protection and water area management. Combining land and underwater data, an integrated digital elevation model is generated, which provides a global perspective for comprehensive analysis. The generated three-dimensional line drawing can intuitively display the terrain features and help identify important terrain changes. Through visual design, the monitoring results are more readable and easy for non-professionals to understand. The three-dimensional digital line drawing finally generated provides a practical tool for dynamic monitoring, which can support timely response and decision-making and enhance the effectiveness of monitoring. Therefore, the present invention improves the accuracy and comprehensiveness of dynamic terrain monitoring through accurate geographic coordinate confirmation, efficient day and night data collection, multi-dimensional model construction, dynamic trend analysis and visual result display.

[0075] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a dynamic terrain monitoring method based on image recognition of the present invention. In this example, the dynamic terrain monitoring method based on image recognition includes the following steps:

[0076] Step S1: Acquire survey area data; confirm the geographic coordinates of the breach on the survey area data to obtain the geographic coordinates of the breach; collect daytime drone oblique photography images on the breach geographic coordinates data to generate daytime breach photography images; perform three-dimensional spatial encryption on the daytime breach photography images to generate daytime breach three-dimensional spatial encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional spatial encrypted data to generate a daytime breach real-scene three-dimensional model;

[0077] In an embodiment of the present invention, relevant data of the survey area are collected, including topographic data, environmental data and historical hydrological data. This can be obtained in the following ways: Remote sensing data: Use satellite images or aerial images for preliminary data collection. Ground survey: Conduct field surveys to collect relevant geographical and topographic information. Database query: Obtain historical measurement and hydrological data from local or national databases. Determine the geographic coordinate system used (such as WGS84 or China National Geodetic Coordinate System) to ensure data consistency. Use GPS equipment or existing geographic information system (GIS) software to locate the breach in the survey area and confirm its accurate geographic coordinate data. Select a specific drone (such as DJI Phantom series or other professional drones) and configure a high-resolution camera. According to the geographic coordinate data of the breach, formulate the flight path and altitude of the drone to ensure coverage of the entire survey area. Fly the drone during the day and perform oblique photography according to the preset path to ensure that images of the breach area are collected at multiple angles and altitudes. Preprocess the collected images, including denoising, contrast enhancement and color correction, to improve image quality. Select a specific 3D space encryption algorithm (such as Structure from Motion (SfM)) to reconstruct 3D point cloud data from multi-view images. Generate daytime breach 3D space encryption data through algorithm processing. Perform point cloud processing on the daytime breach 3D space encryption data, including denoising, downsampling and density adjustment, to obtain fine point cloud data. Generate a 3D model using point cloud data and mesh generation algorithms (such as Poisson Surface Reconstruction). Optimize the generated 3D model, including texture mapping, smoothing and detail enhancement, to ensure that the generated model truly reflects the terrain characteristics of the breach area.

[0078] Step S2: collecting nighttime drone point cloud data according to the breach geographic coordinate data to generate nighttime breach laser point cloud data; measuring the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; constructing a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; performing water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitoring the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data to generate breach land terrain monitoring data;

[0079] In the embodiment of the present invention, by selecting a drone suitable for night operations, equipped with a laser radar (LiDAR) sensor to collect accurate point cloud data. The drone should have the ability to fly at night and have an automatic obstacle avoidance function. Based on the geographic coordinate data of the breach, the flight path of the night drone is formulated. The flight altitude and coverage should ensure complete coverage of the breach and its surrounding areas, especially the key areas of the breach. Fly at night, use the LiDAR sensor to scan the breach area, and collect high-precision laser point cloud data. The collected data should include reflection signals from the ground, vegetation and water surface. Use the real-life three-dimensional model of the daytime breach generated in step S1 to conduct a detailed inspection of the geometric features of the breach. According to the daytime three-dimensional model, the boundary points of the breach mouth are selected, and the width of the mouth is calculated. Accurate distance measurement can be performed through geometric analysis tools to generate daytime breach mouth width data. The laser point cloud data collected at night are preprocessed, including noise removal, ground point extraction, and separation of water surface reflection signals. Through these processing steps, the quality and accuracy of the point cloud data are ensured. Using point cloud data, a digital surface model of the night water area is constructed. DSM can be generated through interpolation algorithms, showing the elevation and changes of the water surface, reflecting the topographic characteristics of the water surface at night. The digital surface model of the night water area is rasterized to convert the continuous water surface elevation data into equally spaced grid cells. Each grid cell represents the water surface elevation value at that location. Elevation trend analysis is performed based on the water surface elevation data of the grid cells. The analysis method includes statistically analyzing the change amplitude, gradient change and trend direction of the water surface elevation, thereby generating the night breach water surface height trend data. The daytime breach mouth width data and the nighttime water surface height trend data are fused and analyzed. By comparing the water surface height changes and mouth widths in different time periods, the change trend of the breach land topography is identified. Using the nighttime water surface elevation trend data and the daytime breach mouth width data, the breach land topography is monitored to detect land settlement, landslide, erosion and other land terrain changes. Finally, the breach land topography monitoring data is generated, which can be used for further terrain analysis and breach management decisions.

[0080] Step S3: Acquire breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a water-land integrated digital elevation model of the breach;

[0081] In an embodiment of the present invention, by deploying flow velocity sensors in the breach area, usually selecting devices such as acoustic current meters (ADCP), the water flow velocity can be monitored in real time. Automatic collection of flow velocity data is performed at set time intervals. Ensure that data collection covers different depths and water flow conditions of the breach so as to obtain comprehensive flow velocity data. The recorded flow velocity data should include timestamps, flow velocity values ​​and water depth information to facilitate subsequent analysis and application. According to the acquired breach flow velocity data and breach geographic coordinate data, a detailed underwater topographic measurement plan is formulated, including measurement range, path and frequency. Use a depth-measuring unmanned vessel (such as an unmanned boat equipped with a sonar system) to perform underwater topography measurement along the measurement path and collect water depth data. The measured water depth data is processed to generate preliminary breach underwater topographic measurement data, including bottom contours and topographic features. During the measurement process, it is necessary to synchronously monitor water level changes and obtain real-time water level data. A water level sensor can be used for real-time monitoring. The collected water depth data is checked with the monitored water level data to perform water level correction. The calculation formula is: in, is the corrected water depth, is the measured water depth, is the current water level. The complete underwater terrain data of the breach area is generated through the depth data after water level correction to reflect the real underwater terrain conditions. A regular underwater terrain monitoring plan is formulated to monitor the changes in the complete underwater terrain data. Unmanned boats or other underwater detection equipment are used regularly to repeat underwater terrain measurements and collect the latest underwater terrain data. The underwater terrain data of different time periods are compared, the terrain change trend is analyzed, the breach underwater terrain monitoring data is generated, and changes such as sedimentation and erosion are identified. The breach land terrain monitoring data and the breach underwater terrain monitoring data generated in the previous steps are fused. Ensure the accuracy and consistency of the two data. Point cloud processing is performed on the fused data, including denoising and filtering outliers. Spatial interpolation technology (such as Kriging interpolation) is applied to generate a continuous digital elevation model. Combining land and underwater data, a complete integrated land and water digital elevation model of the breach is constructed to provide a global view and accurate terrain information to support subsequent analysis and decision-making.

[0082] Step S4: Project the three-dimensional coordinates of the feature line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring of the breach.

[0083] In an embodiment of the present invention, a water-land integrated digital elevation model of the breach and a feature line set (such as contour lines, water body boundary lines, shorelines, and important terrain feature lines) extracted from the model are prepared. Use geographic information system (GIS) software or three-dimensional modeling software (such as ArcGIS, QGIS, or SketchUp, etc.) to perform three-dimensional coordinate projection. Convert the two-dimensional coordinate data (for example, plane coordinates) of the feature line set into three-dimensional coordinate data, and consider the elevation value (Z axis) to achieve accurate three-dimensional representation. Ensure that the feature lines are accurately aligned with the digital elevation model to reflect the real terrain features. Based on the projection results, generate a preliminary three-dimensional line drawing of the breach to show the layout of the feature lines in three-dimensional space. Define the colors and styles of different types of lines according to monitoring needs and actual conditions. For example: Contour lines: Use different shades of color to represent different elevation levels. Water body boundary lines: Use blue lines to distinguish. Shorelines: Use green or black to make them obvious. Important terrain feature lines: Use red or other conspicuous colors. Use the style editing function in 3D modeling software or GIS tools to set colors, line widths, virtual and real lines, transparency, etc. Adjust the presentation of lines to improve the readability and aesthetics of 3D digital line drawings. For example, set shadows, lighting effects, etc. to make the line drawings more three-dimensional visually. Export the 3D line drawings of breaches after style settings into specific formats (such as 3D PDF, GeoTIFF, OBJ, etc.) for easy storage, sharing, and display. Prepare an operating interface for dynamic monitoring to ensure that the 3D digital line drawings can be superimposed and displayed with real-time data (such as water level, flow rate, etc.) to support dynamic updates and monitoring analysis. Combine sensor data with 3D digital line drawings to display real-time changes through GIS or 3D visualization software to facilitate monitoring decisions.

[0084] Preferably, step S1 comprises the following steps:

[0085] Step S11: Acquire survey area data;

[0086] Step S12: confirming the geographical coordinates of the breach on the survey area data to obtain the geographical coordinate data of the breach; collecting daytime drone oblique photography images based on the geographical coordinate data of the breach to generate daytime breach photography images;

[0087] Step S13: performing image preprocessing on the daytime ulcer photography image to generate a standard daytime ulcer photography image, wherein the image preprocessing includes image filtering, image brightness enhancement and image resolution enhancement;

[0088] Step S14: encrypting the standard daytime breach photography image in three-dimensional space to generate daytime breach three-dimensional space encrypted data; constructing a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model.

[0089] In an embodiment of the present invention, relevant geographic information of the survey area is collected, including topography, soil type, historical breach data, meteorological data, etc. GPS and GIS technology are used to record the boundaries and features of the survey area to ensure the accuracy of subsequent data collection. The survey area data is analyzed to identify potential breach areas. Geographic Information System (GIS) software is used to confirm the breach geographic coordinates of these areas and generate breach geographic coordinate data. Based on the confirmed breach geographic coordinate data, the flight path of the drone is planned to ensure that all breach areas are covered. The drone is used to collect daytime oblique photography images to generate high-resolution daytime breach photography images. The collected daytime breach photography images are preprocessed to ensure image quality and consistency. A high-pass filter is used to remove image noise and improve image clarity. Histogram equalization technology is applied to enhance the contrast of the image and make the details more obvious. Super-resolution reconstruction technology is used to improve the resolution of the image so that more details can be obtained in subsequent three-dimensional modeling. Using standard daytime breach photography images, apply 3D reconstruction algorithms (such as Structure from Motion (SfM) or Multi-View Stereo (MVS)) to perform 3D space encryption, generate 3D space encrypted data of daytime breaches, and form a point cloud data set. Use point cloud processing software (such as CloudCompare or Meshlab) to process and optimize the point cloud data and extract surface features. Based on the optimized daytime breach 3D space encrypted data, construct a 3D model and generate a real-life 3D model of the daytime breach for subsequent analysis and visualization.

[0090] Preferably, step S14 comprises the following steps:

[0091] Step S141: performing image segmentation on the standard daytime ulcer photography image to generate a ulcer region segmentation image; performing ulcer feature extraction on the ulcer region segmentation image to obtain daytime ulcer region feature data; converting the daytime ulcer region feature data into data point cloud to generate daytime ulcer region feature point cloud data;

[0092] Step S142: performing point cloud density analysis on the daytime breach area feature point cloud data to generate regional feature point cloud density data; performing data block segmentation on the daytime breach area feature point cloud data according to the regional feature point cloud density data to generate regional feature point cloud segmentation data blocks;

[0093] Step S143: encrypting the regional feature point cloud segmentation data blocks to generate regional feature point cloud encrypted data blocks; performing three-dimensional spatial arrangement confusion on the regional feature point cloud encrypted data blocks to generate three-dimensional spatial encrypted data of the daytime breach; performing three-dimensional gridding on the three-dimensional spatial encrypted data of the daytime breach to generate a preliminary three-dimensional grid model of the daytime breach;

[0094] Step S144: extract color and texture information from the standard daytime breach photography image to obtain daytime breach color data and daytime breach texture information data; perform texture color mapping on the preliminary three-dimensional mesh model of the daytime breach using the daytime breach color data and daytime breach texture information data to generate a real-scene three-dimensional model of the daytime breach.

[0095] In an embodiment of the present invention, a standard daytime ulcer photographic image is processed by using an image segmentation algorithm (such as K-means or U-Net network), the ulcer area is separated, and a segmented image of the ulcer area is generated to highlight the characteristics of the ulcer. The characteristics of the ulcer area, including edges, shapes, sizes, etc., are extracted by image processing technology to generate feature data of the daytime ulcer area. Computer vision technology (such as SIFT or SURF) is used to further enhance the accuracy of the feature data. The extracted feature data of the daytime ulcer area is converted into point clouds to generate feature point cloud data of the daytime ulcer area for subsequent three-dimensional processing. The point cloud density of the feature point cloud data is calculated to generate regional feature point cloud density data, and high-density and low-density areas are identified. The point cloud is carefully analyzed using a density estimation method (such as a Gaussian mixture model) to extract meaningful spatial information. According to the regional feature point cloud density data, the feature point cloud data of the daytime ulcer area is segmented into data blocks to generate regional feature point cloud segmentation data blocks. A voxel grid method or a segmentation algorithm based on regional growth is used to ensure that each data block can represent the corresponding feature area. Encrypt the data blocks of the regional feature point cloud segmentation to protect the privacy and integrity of the data. AES encryption algorithm or other suitable data encryption technology can be used. Perform three-dimensional spatial arrangement confusion on the encrypted data blocks to generate three-dimensional spatial encrypted data of daytime breaches to enhance data security. Use a random arrangement algorithm to reconfigure the data blocks in space to avoid direct leakage of feature information. Perform three-dimensional meshing on the three-dimensional spatial encrypted data of daytime breaches to generate a preliminary three-dimensional mesh model of daytime breaches. Use an algorithm (such as Delaunay triangulation) to mesh the point cloud data to create a high-fidelity three-dimensional mesh. Extract color data and texture information data from standard daytime breach photography images to generate daytime breach color data and daytime breach texture information data. Use texture analysis technology (such as GLCM or LBP) to obtain a detailed texture description. Use the extracted daytime breach color data and texture information data to perform texture color mapping on the preliminary three-dimensional mesh model of daytime breaches. UV mapping technology is used to accurately map color and texture information to the surface of the 3D model to generate the final real-life 3D model of the breach during the day.

[0096] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0097] Step S21: using a laser radar drone to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data;

[0098] Step S22: measuring the opening width of the daytime breach real-scene three-dimensional model to obtain the opening width data of the daytime breach; confirming the fuzzy area of ​​the nighttime breach on the laser point cloud data of the nighttime breach according to the opening width data of the daytime breach to obtain the point cloud data of the fuzzy area of ​​the nighttime breach; constructing a digital model of the surface of the nighttime water area based on the point cloud data of the fuzzy area of ​​the nighttime breach to generate a digital surface model of the nighttime water area;

[0099] Step S23: performing night breach water surface height analysis on the night water area digital surface model to generate night breach water surface height data; performing water surface elevation raster analysis on the night breach water surface height data to generate night breach water surface height trend data;

[0100] Step S24: Monitor the land topography of the breach through the daytime breach gate width data and the nighttime breach water surface height trend data, thereby generating the land topography monitoring data of the breach.

[0101] In an embodiment of the present invention, a laser radar (LiDAR) device is installed on the drone to ensure that it can work at night and is equipped with appropriate lighting and sensors. According to the pre-confirmed breach geographic coordinate data, the flight path of the drone is set to ensure that all target areas are covered. Laser point cloud data collection is performed at night, the laser point cloud data of the breach at night is recorded, and a laser point cloud data set of the breach at night is generated. Real-time data monitoring technology is applied to ensure the quality and accuracy of data collection. The width of the mouth is measured using the real-life three-dimensional model of the breach during the day, and the mouth width data of the breach during the day is extracted. Measuring tools (such as rangefinders or measurement functions in three-dimensional modeling software) are used to ensure the accuracy of the measurement. Based on the daytime mouth width data, the laser point cloud data of the breach at night is analyzed to identify the fuzzy area of ​​the breach at night. The point cloud filtering and classification algorithm is used to separate the clear area and the fuzzy area, and the point cloud data of the fuzzy area of ​​the breach at night is generated. The point cloud data of the fuzzy area of ​​the breach at night is used to construct a digital surface model of the water area at night, and a digital surface model of the water area at night is generated. Use point cloud reconstruction technology (such as Poisson Surface Reconstruction) to generate a high-precision water surface model. Perform water surface height analysis on the digital surface model of the night water area and extract water surface height data. Use digital terrain analysis software (such as ArcGIS or QGIS) to extract and analyze water surface elevation to ensure accuracy. Perform water surface elevation raster analysis based on the night water surface height data to generate night breach water surface height trend data. Use interpolation methods (such as Kriging interpolation or inverse distance weighted method) to generate rasterized water surface elevation data for subsequent analysis. Integrate the daytime breach gate width data and the night breach water surface height trend data in preparation for comprehensive analysis. Based on the integrated data, conduct breach land terrain monitoring, use analysis software to visualize and evaluate terrain changes, generate breach land terrain monitoring data, and record important information such as terrain changes, potential risks, and water level impacts.

[0102] Preferably, step S23 includes the following steps:

[0103] Step S231: performing rough water surface area extraction on the nighttime water area digital surface model based on a preset elevation range to obtain rough water surface area data;

[0104] Step S232: performing regional reflectivity difference calculation on the rough water surface area data to obtain regional reflectivity difference data; performing regional boundary adjustment on the rough water surface area data by using the regional reflectivity difference data, thereby generating night breach water surface area data;

[0105] Step S233: selecting water surface key points from the night breach water surface area data to obtain the night breach water surface area key point data, wherein the water surface key points selected include the breach center area, the breach left area and the breach right area; performing water surface elevation profile height comparison on the night breach water surface area data according to the night breach water surface area key point data to generate night breach water surface height data; performing grid unit conversion on the night breach water surface height data to generate water surface height grid units;

[0106] Step S234: Calculate the local elevation slope of the water surface height grid unit to obtain the local elevation slope data of the water surface; perform curve conversion on the local elevation slope data of the water surface to generate a water surface elevation slope curve; perform trend analysis on the water surface elevation slope curve to generate nighttime breach water surface height trend data.

[0107] In an embodiment of the present invention, a preset elevation range (e.g., the minimum and maximum values ​​of the water surface height) of the nighttime water area is set according to historical data or a topographic map. A geographic information system (GIS) software is used to screen the digital surface model of the nighttime water area, extract the water surface area that meets the preset elevation range, generate rough water surface area data, and record the spatial coordinates and elevation information of these areas. The regional reflectivity of the rough water surface area is calculated using laser radar or image data to obtain reflectivity data. By comparing the reflectivity of the rough water surface area with the surrounding terrain, the regional reflectivity difference is calculated to generate regional reflectivity difference data. According to the regional reflectivity difference data, the boundary of the rough water surface area is adjusted to remove artifacts and noise, and more accurate nighttime breach water surface area data is generated. A boundary smoothing algorithm (e.g., regional growing method) is applied to ensure that the boundary of the water surface area is smooth and meets the actual terrain characteristics. In the nighttime breach water surface area data, the breach center area, the left area, and the right area are selected as key points. Ensure that the selected key points can reflect the changes and characteristics of the water surface height. Using the selected water surface key point data, generate a water surface elevation profile and compare the water surface heights of each area. Record the profile data and generate the night breach water surface height data. Convert the night breach water surface height data into a grid unit format for subsequent analysis and visualization. Calculate the local elevation slope of the water surface height grid unit to obtain the local elevation slope data of the water surface. Use terrain analysis tools (such as ArcGIS's elevation analysis tool) to calculate the elevation difference between adjacent grid units and obtain the slope value. Convert the local elevation slope data into a curve to generate a water surface elevation slope curve to facilitate observation of the slope change trend. Apply a smoothing algorithm (such as the Savitzky-Golay filter) to remove high-frequency noise and improve the readability of the curve. Perform a trend analysis on the water surface elevation slope curve to identify the change pattern of the water surface height and generate the night breach water surface height trend data. Use statistical analysis methods (such as linear regression) to evaluate the change trend of the water surface height over time and provide a basis for future water level prediction.

[0108] Preferably, step S232 includes the following steps:

[0109] Step S2321: performing Sobel operator calculation on the rough water surface area data to generate local reflectivity gradient data of the rough water surface area; marking high gradient areas on the local reflectivity gradient data of the rough water surface area to generate high gradient area data of rough water surface reflectivity;

[0110] Step S2322: constructing a gradient enhancement map for the local reflectivity gradient data of the rough water surface area according to the rough water surface reflectivity high gradient area data, generating a water surface area reflectivity gradient enhancement map; performing adjacent gradient difference calculation on the water surface area reflectivity gradient enhancement map, obtaining regional reflectivity difference data;

[0111] Step S2323: performing gradient peak tracking on the rough water surface area data through the regional reflectivity difference data to generate regional gradient peak data; performing boundary smoothing on the rough water surface area data through the regional gradient peak data to generate boundary adjustment data for the breach water surface area at night;

[0112] Step S2324: Perform regional consistency verification on the rough water surface area data based on the nighttime breach water surface area boundary adjustment data, thereby generating nighttime breach water surface area data.

[0113] In the embodiment of the present invention, the Sobel operator is used to perform edge detection on the rough water surface area data to calculate the gradient value of the local reflectivity. The gradient calculation formula is: in and Represent the gradients in the horizontal and vertical directions respectively. The calculated gradient values ​​are combined into local reflectivity gradient data to reflect the changes in the water surface area. According to the set gradient threshold, the high gradient area in the local reflectivity gradient data is identified to generate the rough water surface reflectivity high gradient area data. The regional growing algorithm or threshold segmentation method is used to ensure that the high gradient area is clearly distinguished from the surrounding area. According to the set gradient threshold, the high gradient area in the local reflectivity gradient data is identified to generate the rough water surface reflectivity high gradient area data. The regional growing algorithm or threshold segmentation method is used to ensure that the high gradient area is clearly distinguished from the surrounding area. The gradient difference between adjacent pixels of the water surface area reflectivity gradient enhancement map is calculated to obtain the regional reflectivity difference data. This process can use convolution operations to calculate the changes between adjacent pixels to identify water surface features and boundaries. The regional reflectivity difference data is used to track the gradient peak of the rough water surface area to generate regional gradient peak data. The peak detection algorithm (such as local maximum detection) is applied to determine the significant water surface feature points. The boundary of the rough water surface area is smoothed by the regional gradient peak data to eliminate noise and artifacts. Smoothing algorithms (such as Gaussian filtering or median filtering) are used to ensure the continuity and naturalness of the boundaries and generate boundary adjustment data for the night breach water surface area. Based on the boundary adjustment data for the night breach water surface area, the consistency of the rough water surface area data is verified. A consistency metric (such as the Jaccard index or the Dice coefficient) is used to evaluate the overlap between the adjusted boundary and the original area data. If the consistency verification passes, the boundary adjustment data for the night breach water surface area is integrated to generate the final night breach water surface area data. Ensure that the generated data can accurately reflect the actual situation of the water surface and provide a basis for subsequent hydrological monitoring and analysis.

[0114] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0115] Step S31: obtaining breach flow velocity data;

[0116] Step S32: laying out the underwater topography survey plan line based on the breach flow velocity data and the breach geographic coordinate data, and generating an underwater topography survey plan line layout map; performing underwater topography survey of the breach on the underwater topography survey plan line layout map using the cross-section method based on the depth-measuring unmanned ship, and generating the breach underwater topography survey data;

[0117] Step S33: perform water depth verification on the breach underwater topography measurement data to generate breach underwater topography verification data; perform water level correction on the breach underwater topography verification data to generate complete underwater topography data of the breach area;

[0118] Step S34: Perform underwater terrain monitoring on the complete underwater terrain data of the breach area to generate underwater terrain monitoring data of the breach; construct a digital elevation model using the breach land terrain monitoring data and the breach underwater terrain monitoring data to generate a land and water integrated digital elevation model of the breach.

[0119] In the embodiment of the present invention, by selecting a specific flow velocity measuring instrument, such as an acoustic Doppler current meter (ADCP) or a flow velocity sensor. Flow velocity monitoring points are set in the breach area to collect flow velocity data regularly. The data should include a time series record of the flow velocity and record relevant environmental factors (such as water level, meteorological conditions, etc.). The acquired original flow velocity data is subjected to denoising, the average flow velocity and the flow velocity change rate are calculated, and the breach flow velocity data that can be used for subsequent analysis are generated. According to the breach flow velocity data and the breach geographic coordinate data, the plan line for underwater topography measurement is designed. These measurement lines should cover important areas to ensure that the underwater topographic features can be reflected, and a layout diagram of the underwater topographic measurement plan line is generated, and the coordinate information of each measurement point is marked. A depth measuring unmanned ship (such as an unmanned ship with sonar equipment) is used to perform underwater topography measurement along the laid measurement line. The cross-section method is used to record water depth data at a certain distance to generate breach underwater topography measurement data. The obtained breach underwater topography measurement data is subjected to water depth verification. The measurement results are verified using known water level data. If outliers or errors are found, use statistical analysis methods to correct the data and generate verification data for underwater topography of the breach. According to the changes in water levels, correct the verified water depth data, consider the impact of factors such as tides and rainfall on the water level, and generate complete underwater topography data for the breach area for subsequent analysis and application. Monitor the underwater topography of the breach area regularly, and use new underwater topography measurement data to update the complete underwater topography data. Record changes in underwater topography and generate monitoring data for underwater topography of the breach. Use the acquired onshore and underwater topography monitoring data of the breach and combine them with geographic information system (GIS) technology to build a digital elevation model (DEM). Integrate underwater and onshore topography data through interpolation methods (such as Kriging interpolation, inverse distance weighted method, etc.) to generate a water-land integrated digital elevation model of the breach.

[0120] Preferably, constructing a digital elevation model using breach land terrain monitoring data and breach underwater terrain monitoring data includes:

[0121] Perform point cloud conversion on the breach land terrain monitoring data to generate breach land terrain monitoring point cloud data; perform ground attachment filtering on the breach land terrain monitoring point cloud data to generate breach land terrain monitoring filtered data, wherein the ground attachment filtering includes vegetation, water surface and vehicles;

[0122] Extract spatial three-dimensional features from the breach underwater terrain monitoring data and breach land terrain monitoring filtered data to obtain breach land terrain spatial feature data and breach underwater terrain spatial feature data; select spatial common points from the breach land terrain spatial feature data and breach underwater terrain spatial feature data to obtain land and underwater spatial boundary point data;

[0123] The vertex elevations of the onshore and underwater spatial boundary point data are triangulated to generate onshore and underwater triangular meshes; polygonal boundary overlap adjustment is performed on the onshore and underwater triangular meshes to generate onshore and underwater triangular mesh adjustment data; spatial interpolation is performed on the onshore and underwater triangular meshes based on the onshore and underwater triangular mesh adjustment data to generate a land and water integrated digital elevation model of the breach.

[0124] In an embodiment of the present invention, the elevation data of the measuring point is extracted from the breach land terrain monitoring data. The spatial coordinate (X, Y, Z) information is used to convert the monitoring data into point cloud data to generate breach land terrain monitoring point cloud data. A threshold-based method, morphological operation or machine learning algorithm (such as a deep learning network) is used to identify ground attachments (such as vegetation, water surface, vehicles, etc.). The monitoring point cloud data is filtered to remove the identified ground attachments to generate breach land terrain monitoring filtered data. PCA (principal component analysis), normal estimation or curvature calculation methods are used to extract spatial three-dimensional features from the breach land and underwater terrain monitoring data, respectively, to generate breach land terrain spatial feature data and breach underwater terrain spatial feature data for subsequent processing. A similarity matching algorithm (such as an ICP algorithm) is used to compare the land and underwater spatial feature data, identify common points that overlap in space, and generate land and underwater spatial intersection point data as the basis for subsequent triangulation. According to the data of the onshore and underwater spatial boundary points, the Delaunay triangulation algorithm is used to triangulate the boundary points to generate a preliminary triangular mesh. The elevation data of each vertex of the triangular mesh is calculated to generate the onshore and underwater triangular meshes. The boundary overlap areas of the onshore and underwater triangular meshes are detected, and the overlapping polygon boundaries are identified. The overlapping boundaries are smoothed and adjusted to eliminate discontinuities and generate the onshore and underwater triangular mesh adjustment data. According to the data characteristics, a specific spatial interpolation method is selected, such as Kriging interpolation, IDW (inverse distance weighted) or spline interpolation. Based on the onshore and underwater triangular mesh adjustment data, the onshore and underwater terrains are spatially interpolated to generate a smooth digital elevation model.

[0125] Preferably, step S4 comprises the following steps:

[0126] Step S41: extracting contour lines from the breach water-land integrated digital elevation model to obtain the breach terrain contour lines; extracting water systems and shorelines from the breach waterway integrated digital elevation model to obtain water body boundary lines and breach shoreline layers;

[0127] Step S42: extracting important terrain feature lines from the breach water-land integrated digital elevation model to generate a breach important terrain feature line set, wherein the important terrain feature lines extracted include breach edge lines, steep slope area boundary lines, depression lines, and ridge lines;

[0128] Step S43: Perform three-dimensional coordinate projection on the breach terrain contour lines, water body boundary lines, breach shoreline layers, and important terrain feature line sets of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring operations on the breach.

[0129] In an embodiment of the present invention, by selecting a specific contour extraction algorithm, such as Marching Squares or Terrain Analysis method. The integrated digital elevation model of the breach is processed to extract contour lines with a certain spacing, and a contour map of the breach terrain is generated. A water flow analysis algorithm (such as the D8 flow direction algorithm) is used to extract the water flow path from the digital elevation model and generate the water body boundary line. According to the relationship between the water level line and the terrain elevation data, the shoreline of the breach is extracted and a shoreline layer is generated. The digital elevation model is edge detected using the Canny edge detection or the Sobel operator to extract the edge line of the breach. The slope map is calculated to determine the area with significant slope changes and the boundary line is extracted. By analyzing the concave and convex points of the terrain (the lowest point and the highest point), the depression line and the ridge line are extracted to generate a set of important terrain feature lines. The terrain contour lines of the breach, the water body boundary line, the shoreline layer and the set of important terrain feature lines are projected into three-dimensional coordinates to generate a three-dimensional line drawing of the breach. GIS software or a custom script (such as Python's matplotlib library) is used for three-dimensional coordinate conversion and visualization. The generated 3D line drawing is styled, including line color, line type (such as solid line, dashed line) and width settings to improve visibility and differentiation. Different colors and styles are set according to the importance of feature lines to make important terrain features more prominent.

[0130] In this specification, a dynamic terrain monitoring system based on image recognition is provided, which is used to execute the above-mentioned dynamic terrain monitoring method based on image recognition. The dynamic terrain monitoring system based on image recognition includes:

[0131] The daytime land measurement module is used to obtain survey area data; confirm the breach geographic coordinates of the survey area data to obtain the breach geographic coordinates data; collect the breach geographic coordinates data by daytime drone oblique photography to generate daytime breach photography images; encrypt the daytime breach photography images in three-dimensional space to generate daytime breach three-dimensional space encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model;

[0132] The nighttime land measurement module is used to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data; measure the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; construct a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; perform water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitor the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data, thereby generating breach land terrain monitoring data;

[0133] The land-water integrated monitoring module is used to obtain the breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a land-water integrated digital elevation model of the breach;

[0134] The dynamic terrain monitoring module is used to perform three-dimensional coordinate projection of the characteristic line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line drawing of the breach; set the line color and style of the three-dimensional line drawing of the breach to generate a three-dimensional digital line drawing of the breach to perform dynamic terrain monitoring operations of the breach.

[0135] The beneficial effect of the present invention is that by confirming the geographic coordinates of the breach in the survey area data, the data obtained is ensured to be highly accurate, laying the foundation for subsequent monitoring and analysis. The daytime breach photography images generated by using drones for daytime oblique photography can provide rich surface information and support the subsequent three-dimensional model construction. The three-dimensional model is constructed based on the three-dimensional spatial encrypted data of the daytime breach, and the generated daytime breach real-life three-dimensional model can intuitively display the terrain characteristics of the breach area, providing a basis for further analysis. According to the breach geographic coordinate data, nighttime drone point cloud data is collected, and the generated nighttime breach laser point cloud data can capture detailed information about the nighttime water area, supplementing the shortcomings of daytime collection. The mouth width of the daytime real-life three-dimensional model is measured, and the generated mouth width data helps to evaluate the degree of opening of the breach and the dynamics of water flow. The digital surface model of the nighttime water area is generated based on the nighttime laser point cloud data, and the water surface state and its changes can be analyzed. Through the water surface elevation raster analysis, the generated nighttime breach water surface height trend data provides necessary data support for hydrological analysis and flow rate evaluation. The breach land terrain monitoring data generated by integrating the daytime breach gate width data and nighttime water surface height trend data can effectively reflect the terrain changes in the breach area. Obtaining the breach flow velocity data provides necessary parameters for underwater terrain measurement, which helps to accurately evaluate the flow characteristics. Underwater terrain measurement is performed based on the breach flow velocity data and geographic coordinates, and the generated measurement data provides a basis for subsequent water level correction and monitoring. After water level correction, the generated complete underwater terrain data provides high-quality data support for underwater feature analysis. Combining the onshore and underwater terrain monitoring data, the generated digital elevation model can fully reflect the water-land boundary characteristics of the breach area. The three-dimensional coordinate projection of the feature line set of the breach water-land integrated digital elevation model is performed, and the generated three-dimensional line drawing can intuitively display the terrain characteristics and is easy to understand. By setting the line color and style, the generated three-dimensional digital line drawing of the breach is more visually intuitive, which is helpful for performing dynamic terrain monitoring operations of the breach. This systematic monitoring process provides decision makers with efficient information support through detailed data collection and analysis, which helps to quickly respond to potential terrain changes or disaster risks. The generated three-dimensional digital line map can not only be used for water conservancy project monitoring, but also provide support for geological disaster assessment, environmental protection and infrastructure construction, and has broad application prospects. Therefore, the present invention improves the accuracy and comprehensiveness of dynamic terrain monitoring through precise geographic coordinate confirmation, efficient day and night data collection, multi-dimensional model construction, dynamic trend analysis and visual result display.

[0136] 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.

[0137] 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 dynamic terrain monitoring method based on image recognition, characterized in that: The following steps are involved: Step S1: Acquire survey area data; confirm the breach geographic coordinates of the survey area data to obtain the breach geographic coordinates data; collect daytime drone oblique photography images of the breach geographic coordinates data to generate daytime breach photography images; Perform three-dimensional space encryption on the daytime breach photography image to generate three-dimensional space encryption data of the daytime breach; construct a three-dimensional model based on the three-dimensional space encryption data of the daytime breach to generate a three-dimensional model of the real scene of the daytime breach; Step S2: collecting nighttime drone point cloud data according to the breach geographic coordinate data to generate nighttime breach laser point cloud data; measuring the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; constructing a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; performing water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitoring the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data to generate breach land terrain monitoring data; Step S3: Acquire breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a water-land integrated digital elevation model of the breach; Step S4: Project the three-dimensional coordinates of the feature line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring of the breach.

2. The dynamic terrain monitoring method based on image recognition according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire survey area data; Step S12: confirming the geographical coordinates of the breach on the survey area data to obtain the geographical coordinate data of the breach; collecting daytime drone oblique photography images based on the geographical coordinate data of the breach to generate daytime breach photography images; Step S13: performing image preprocessing on the daytime ulcer photography image to generate a standard daytime ulcer photography image, wherein the image preprocessing includes image filtering, image brightness enhancement and image resolution enhancement; Step S14: encrypting the standard daytime breach photography image in three-dimensional space to generate daytime breach three-dimensional space encrypted data; constructing a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model.

3. The dynamic terrain monitoring method based on image recognition according to claim 2 is characterized in that: Step S14 includes the following steps: Step S141: performing image segmentation on the standard daytime ulcer photography image to generate a ulcer region segmentation image; performing ulcer feature extraction on the ulcer region segmentation image to obtain daytime ulcer region feature data; converting the daytime ulcer region feature data into data point cloud to generate daytime ulcer region feature point cloud data; Step S142: performing point cloud density analysis on the daytime breach area feature point cloud data to generate regional feature point cloud density data; performing data block segmentation on the daytime breach area feature point cloud data according to the regional feature point cloud density data to generate regional feature point cloud segmentation data blocks; Step S143: encrypting the regional feature point cloud segmentation data blocks to generate regional feature point cloud encrypted data blocks; performing three-dimensional spatial arrangement confusion on the regional feature point cloud encrypted data blocks to generate three-dimensional spatial encrypted data of the daytime breach; performing three-dimensional gridding on the three-dimensional spatial encrypted data of the daytime breach to generate a preliminary three-dimensional grid model of the daytime breach; Step S144: extract color and texture information from the standard daytime breach photography image to obtain daytime breach color data and daytime breach texture information data; perform texture color mapping on the preliminary three-dimensional mesh model of the daytime breach using the daytime breach color data and daytime breach texture information data to generate a real-scene three-dimensional model of the daytime breach.

4. The dynamic terrain monitoring method based on image recognition according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: using a laser radar drone to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data; Step S22: measuring the opening width of the daytime breach real-scene three-dimensional model to obtain the opening width data of the daytime breach; confirming the fuzzy area of ​​the nighttime breach on the laser point cloud data of the nighttime breach according to the opening width data of the daytime breach to obtain the point cloud data of the fuzzy area of ​​the nighttime breach; constructing a digital model of the surface of the nighttime water area based on the point cloud data of the fuzzy area of ​​the nighttime breach to generate a digital surface model of the nighttime water area; Step S23: performing night breach water surface height analysis on the night water area digital surface model to generate night breach water surface height data; performing water surface elevation raster analysis on the night breach water surface height data to generate night breach water surface height trend data; Step S24: Monitor the land topography of the breach through the daytime breach gate width data and the nighttime breach water surface height trend data, thereby generating the land topography monitoring data of the breach.

5. The method for dynamic terrain monitoring based on image recognition according to claim 4, characterized in that: Step S23 includes the following steps: Step S231: performing rough water surface area extraction on the nighttime water area digital surface model based on a preset elevation range to obtain rough water surface area data; Step S232: performing regional reflectivity difference calculation on the rough water surface area data to obtain regional reflectivity difference data; performing regional boundary adjustment on the rough water surface area data by using the regional reflectivity difference data, thereby generating night breach water surface area data; Step S233: selecting water surface key points from the night breach water surface area data to obtain the night breach water surface area key point data, wherein the water surface key points selected include the breach center area, the breach left area and the breach right area; performing water surface elevation profile height comparison on the night breach water surface area data according to the night breach water surface area key point data to generate night breach water surface height data; performing grid unit conversion on the night breach water surface height data to generate water surface height grid units; Step S234: Calculate the local elevation slope of the water surface height grid unit to obtain the local elevation slope data of the water surface; perform curve conversion on the local elevation slope data of the water surface to generate a water surface elevation slope curve; perform trend analysis on the water surface elevation slope curve to generate nighttime breach water surface height trend data.

6. The method for dynamic terrain monitoring based on image recognition according to claim 5, characterized in that: Step S232 includes the following steps: Step S2321: performing Sobel operator calculation on the rough water surface area data to generate local reflectivity gradient data of the rough water surface area; marking high gradient areas on the local reflectivity gradient data of the rough water surface area to generate high gradient area data of rough water surface reflectivity; Step S2322: constructing a gradient enhancement map for the local reflectivity gradient data of the rough water surface area according to the rough water surface reflectivity high gradient area data, generating a water surface area reflectivity gradient enhancement map; performing adjacent gradient difference calculation on the water surface area reflectivity gradient enhancement map, obtaining regional reflectivity difference data; Step S2323: performing gradient peak tracking on the rough water surface area data through the regional reflectivity difference data to generate regional gradient peak data; performing boundary smoothing on the rough water surface area data through the regional gradient peak data to generate boundary adjustment data for the breach water surface area at night; Step S2324: Perform regional consistency verification on the rough water surface area data based on the nighttime breach water surface area boundary adjustment data, thereby generating nighttime breach water surface area data.

7. The method for dynamic terrain monitoring based on image recognition according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: obtaining breach flow velocity data; Step S32: laying out the underwater topography survey plan line based on the breach flow velocity data and the breach geographic coordinate data, and generating an underwater topography survey plan line layout map; performing underwater topography survey of the breach on the underwater topography survey plan line layout map using the cross-section method based on the depth-measuring unmanned ship, and generating the breach underwater topography survey data; Step S33: perform water depth verification on the breach underwater topography measurement data to generate breach underwater topography verification data; perform water level correction on the breach underwater topography verification data to generate complete underwater topography data of the breach area; Step S34: Perform underwater terrain monitoring on the complete underwater terrain data of the breach area to generate underwater terrain monitoring data of the breach; construct a digital elevation model using the breach land terrain monitoring data and the breach underwater terrain monitoring data to generate a land and water integrated digital elevation model of the breach.

8. The method for dynamic terrain monitoring based on image recognition according to claim 7, characterized in that: The construction of digital elevation model using breach land terrain monitoring data and breach underwater terrain monitoring data includes: Perform point cloud conversion on the breach land terrain monitoring data to generate breach land terrain monitoring point cloud data; perform ground attachment filtering on the breach land terrain monitoring point cloud data to generate breach land terrain monitoring filtered data, wherein the ground attachment filtering includes vegetation, water surface and vehicles; Extract spatial three-dimensional features from the breach underwater terrain monitoring data and breach land terrain monitoring filtered data to obtain breach land terrain spatial feature data and breach underwater terrain spatial feature data; select spatial common points from the breach land terrain spatial feature data and breach underwater terrain spatial feature data to obtain land and underwater spatial boundary point data; The vertex elevations of the onshore and underwater spatial boundary point data are triangulated to generate onshore and underwater triangular meshes; polygonal boundary overlap adjustment is performed on the onshore and underwater triangular meshes to generate onshore and underwater triangular mesh adjustment data; spatial interpolation is performed on the onshore and underwater triangular meshes based on the onshore and underwater triangular mesh adjustment data to generate a land and water integrated digital elevation model of the breach.

9. The method for dynamic terrain monitoring based on image recognition according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: extracting contour lines from the breach water-land integrated digital elevation model to obtain the breach terrain contour lines; extracting water systems and shorelines from the breach waterway integrated digital elevation model to obtain water body boundary lines and breach shoreline layers; Step S42: extracting important terrain feature lines from the breach water-land integrated digital elevation model to generate a breach important terrain feature line set, wherein the important terrain feature lines extracted include breach edge lines, steep slope area boundary lines, depression lines, and ridge lines; Step S43: Perform three-dimensional coordinate projection on the breach terrain contour lines, water body boundary lines, breach shoreline layers, and important terrain feature line sets of the breach to generate a three-dimensional line map of the breach; set the line color and style of the three-dimensional line map of the breach to generate a three-dimensional digital line map of the breach to perform dynamic terrain monitoring operations on the breach.

10. A dynamic terrain monitoring system based on image recognition, characterized in that: Used to execute the dynamic terrain monitoring method based on image recognition as claimed in claim 1, the dynamic terrain monitoring system based on image recognition comprises: The daytime land measurement module is used to obtain survey area data; confirm the breach geographic coordinates of the survey area data to obtain the breach geographic coordinates data; collect the breach geographic coordinates data by daytime drone oblique photography to generate daytime breach photography images; encrypt the daytime breach photography images in three-dimensional space to generate daytime breach three-dimensional space encrypted data; construct a three-dimensional model based on the daytime breach three-dimensional space encrypted data to generate a daytime breach real-scene three-dimensional model; The nighttime land measurement module is used to collect nighttime drone point cloud data based on the breach geographic coordinate data to generate nighttime breach laser point cloud data; measure the mouth width of the daytime breach real-life 3D model to obtain the daytime breach mouth width data; construct a nighttime water surface digital model based on the nighttime breach laser point cloud data to generate a nighttime water digital surface model; perform water surface elevation raster analysis on the nighttime water digital surface model to generate nighttime breach water surface height trend data; monitor the breach land terrain through the daytime breach mouth width data and the nighttime breach water surface height trend data, thereby generating breach land terrain monitoring data; The land-water integrated monitoring module is used to obtain the breach flow velocity data; perform underwater topography measurement of the breach based on the breach flow velocity data and the breach geographic coordinate data to generate underwater topography measurement data of the breach; perform water level correction on the underwater topography measurement data of the breach to generate complete underwater topography data of the breach area; perform underwater topography monitoring on the complete underwater topography data of the breach area to generate underwater topography monitoring data of the breach; construct a digital elevation model using the breach land topography monitoring data and the breach underwater topography monitoring data to generate a land-water integrated digital elevation model of the breach; The dynamic terrain monitoring module is used to perform three-dimensional coordinate projection of the characteristic line set on the integrated water and land digital elevation model of the breach to generate a three-dimensional line drawing of the breach; set the line color and style of the three-dimensional line drawing of the breach to generate a three-dimensional digital line drawing of the breach to perform dynamic terrain monitoring operations of the breach.

Images