Desilting geological detection method for inland river and offshore water

Abstract

The invention discloses a desilting geological detection method for inland river and offshore water, and relates to the technical field of geological detection, the geological detection steps are as follows: S1, using an unmanned aerial vehicle to carry a laser radar to fly in a to-be-detected area, and obtaining underwater topographic information data of the to-be-detected area through two-dimensional point-by-point scanning; detecting depth data in cooperation with a bottom mud detection rod, and constructing a three-dimensional image after preprocessing; and S2, based on an edge detection method, processing the constructed three-dimensional image elevation data, calculating the gradient of each pixel point, setting a gradient threshold to determine edge points, connecting the edge points to form a contour, and identifying the complex terrain position in the to-be-detected area. According to the method, the key dredging position can be judged and processed, accurate implementation of the dredging project can be ensured, the dredging efficiency is improved, resource waste is reduced, the benefits of the dredging project are played to the maximum extent, the underwater complex terrain area is effectively detected, the detection precision is improved, and decisions are provided.

Description

Technical Field

[0001] The present invention relates to the field of geological exploration technology, in particular to a geological exploration method for dredging in inland rivers and offshore water bodies. Background Art

[0002] Inland rivers and offshore waters play a vital role in regional ecosystems and human economic activities. As important veins of land water systems, inland rivers undertake key functions of irrigation, transportation, and regional climate regulation. Their ecological health directly affects agricultural production, industrial water use, and the quality of life of residents in surrounding areas. Offshore waters are connected to the vast ocean and are an important breeding ground for marine fishery resources. They are also the core areas of maritime transportation and marine energy development economic activities. However, with the rapid development of the economy and the increasing frequency of human activities, inland rivers and offshore waters are facing serious siltation problems. The siltation phenomenon is It is widely present in inland rivers and offshore waters. In inland rivers, large amounts of sediment, industrial waste, solid impurities in domestic sewage, and aquatic debris are continuously deposited, causing the river channel to become shallower and the water flow to become stagnant. This seriously weakens the navigability of inland rivers, increases the risk of flooding, destroys the original aquatic ecological balance, and leads to a reduction in biodiversity. In offshore waters, coastal erosion, rivers carrying sediment into the sea, and marine engineering activities have caused serious sedimentation in ports and waterways, hindering the normal entry and exit of ships, reducing the operating efficiency of ports, and causing significant damage to the marine ecological environment.

[0003] To effectively solve the problem of siltation, dredging projects have become an inevitable choice. Accurately understanding the geological conditions of the dredging area is a key prerequisite for ensuring that dredging projects are carried out efficiently, safely and economically. At present, existing dredging geological detection methods have many limitations. When using seismic wave detection, when seismic waves propagate on the bottom of the water, due to the complex interface between the water body and the stratum, multiple reflections and refractions will occur, resulting in a lot of signal interference, affecting the accurate imaging of deep geological structures, and making it difficult to accurately identify some thin strata or small-scale geological structures. Sonar detection, although sonar can quickly obtain underwater topography information, has limited ability to distinguish the locations of some areas with complex underwater structures. The detection depth and accuracy of geological radar detection are greatly affected by the conductivity of the water body and the uniformity of the underwater medium. In highly conductive water bodies or areas with a large amount of metal debris on the bottom, the radar signal attenuates severely, making image interpretation difficult. There is no effective geological detection method for complex underwater terrain structures, which makes it impossible to collect accurate dredging geological information when facing complex terrain conditions. In this regard, we propose a dredging geological detection method for inland rivers and near-shore waters. Summary of the Invention

[0004] In order to solve the above technical problems, a dredging geological detection method for inland rivers and near-shore water bodies is provided. This technical solution solves the above-mentioned problem of being unable to collect accurate dredging geological information when facing complex terrain conditions.

[0005] To achieve the above objectives, the present invention adopts a technical solution: a geological survey method for dredging in inland rivers and offshore water bodies, wherein the geological survey steps are as follows:

[0006] S1. Use a drone equipped with a laser radar to fly over the area to be detected, and obtain underwater terrain information data of the area to be detected point by point through two-dimensional scanning. Use a bottom mud detection rod to detect depth data, and construct a three-dimensional image after pre-processing.

[0007] S2. Process the constructed 3D image elevation data based on edge detection methods, calculate the gradient of each pixel, set a gradient threshold to determine the edge points, connect the edge points to form a contour, and identify the complex terrain locations in the area to be detected;

[0008] S3. Extract data characteristic values ​​related to complex terrain locations from the three-dimensional model. Based on the preliminary terrain data reports of inland rivers and offshore water bodies obtained by the current geological bureau, construct a statistical model to calculate the average elevation range and area distribution interval statistical parameters of normal complex terrain areas. Compare the characteristic values ​​of each identified complex terrain location with the statistical parameters to identify whether there are any abnormalities.

[0009] S4. Use conventional methods to detect geological data at flat locations in the area to be detected. For the identified complex terrain locations, use underwater robots to detect and obtain underwater silt thickness data and underwater environment data. Based on the needs of the dredging project, establish a dredging threshold range, determine the key dredging locations, and carry out dredging treatment at the key dredging locations.

[0010] Preferably, in step S1, historical geological data is collected in advance to understand the geological background and plan the UAV route. During the flight, the lidar emits a laser beam at a high frequency, which is directed to the river surface and received. The UAV obtains underwater terrain data in a two-dimensional point-by-point scanning manner according to the preset route; the bottom mud detection rod is equipped with a pressure sensor to monitor the water pressure, and the depth of the detection rod is calculated based on the correspondence between water pressure and depth. Each preset point is detected more than three times; the acquired data is preprocessed, and the preprocessing includes data cleaning, data denoising and normalization; the preprocessed data is input into the modeling software to construct a three-dimensional model.

[0011] Preferably, in step S2, for each pixel point in the three-dimensional image elevation data, gradient calculation is performed, and the calculation steps are: first, calculating the gradient of the pixel point in the horizontal and vertical directions, combining them to obtain the total gradient, and finally calculating the direction data of the gradient;

[0012] The three-dimensional image elevation data is a two-dimensional matrix Z(x,y), where x and y are the coordinates of the pixel point, and Z(x,y) represents the elevation value of the pixel point; calculate the horizontal gradient G x , calculated using the central difference formula, the expression is:

[0013]

[0014] Among them G x (x,y) means that in the x direction, the gradient of the current pixel is approximately half the difference between the elevation values ​​of the two adjacent pixels;

[0015] Calculate the vertical gradient G y , also using the central difference formula, the total gradient is obtained by combining the horizontal and vertical gradients, and the combined expression is:

[0016]

[0017] Among them, G(x,y) calculates the gradient size of each pixel point, reflecting the degree of change of the elevation value around the point. y (x, y) is in the y direction, and the gradient of the current pixel is approximately half of the difference in elevation between the two adjacent pixels; the direction of the gradient is calculated as θ, and the expression for the direction of the gradient is:

[0018]

[0019] Where θ(x,y) is the direction of the gradient at each pixel, in radians.

[0020] Preferably, in step S2, a gradient threshold is obtained based on experimental data. When the gradient value of the current pixel is greater than the threshold, the current pixel is determined to be an edge point. After the edge point is determined, the discrete edge points are connected to form a contour. Based on the edge tracking algorithm, pixels with gradient directions similar to the gradient direction of the current point and which are also edge points are found in the neighborhood of the point, and the pixel points are connected in sequence to form a closed contour. By analyzing the contour shape, area and direction characteristics, the location of complex terrain in the area to be detected is identified.

[0021] Preferably, the edge points of similar points are determined based on the calculated pixel gradient values. A pixel is determined to be an edge point when the following conditions are met:

[0022]

[0023] Edge(x,y) is a binary function, a value of 1 indicates that the pixel is an edge point, a value of 0 indicates that it is not an edge point, and T is the gradient threshold obtained;

[0024] Connect edge points to form contours by using the neighborhood N(x i ,y i ) to find other edge points (x j ,y j ), the connection is made when the following conditions are met: the neighborhood point is an edge point, that is, Edge(x j ,y j )=1; the gradient directions are similar. Let the threshold of similar gradient directions be Δθ, then it must satisfy:

[0025] |θ(x i y i )-θ(x j ,y j )|<Δθ

[0026] By continuously starting from an edge point, searching for edge points that meet the conditions in its neighborhood and connecting them, a closed contour is finally formed. For the closed contour formed Define the eigenvalues ​​to describe:

[0027] Contour area It is approximated by counting the number of pixels enclosed by the contour;

[0028] Profile shape factor This is measured using circularity, which is defined as: in is the perimeter of the contour;

[0029] Contour direction The main direction of the contour is determined by analyzing the coordinate change trend of each point on the contour and calculating it;

[0030] By analyzing the features, the location of complex terrain is identified; when the contour area is small, the shape factor deviates greatly from 1, and the direction is tortuous, the area surrounded by the contour is complex terrain, and a comprehensive judgment function is defined when The area enclosed by the contour is considered complex terrain when the following conditions are met:

[0031]

[0032] Here, τ is a comprehensive judgment threshold, which is determined based on the specific application scenario and data characteristics.

[0033] Preferably, in step S3, the model is constructed by selecting polynomial regression as the model algorithm, and the constructed model is trained and optimized. The average elevation is calculated as:

[0034]

[0035] in is the average elevation, h i is the elevation value of the i-th sampling point, and n is the total number of sampling points;

[0036] The formula for calculating the elevation range is:

[0037] R=h max -h min

[0038] where h max is the maximum elevation in the area, h min is the minimum elevation value in the area;

[0039] The statistics of area distribution interval are calculated by using geometric formulas. If it is an irregular area, it is calculated based on the grid method. The area is divided into several small grids. The area of ​​each grid is known. The number of grids contained in the area is counted, and then the total area is multiplied by the area of ​​a single grid. The area distribution interval first determines the grouping interval of the area, and then counts the number of sub-areas falling in each interval, calculates the frequency of each interval, and obtains the statistical parameters of the area distribution interval.

[0040] Preferably, the comparison method in step S3 is based on the Euclidean distance method to perform comparison and judgment to identify whether there is complex terrain with abnormal judgment in the three-dimensional model. If there is a misjudgment of complex terrain, steps S1 and S2 are repeated to continue construction and identification, and abnormal judgment is performed again until there is no abnormality.

[0041] Preferably, the conventional methods in step S4 include geological drilling, geological radar and electrical exploration, and geological information is detected in a flat position; in the acquired complex terrain, detection is carried out based on an underwater robot, wherein the silt thickness data is obtained through sonar measurement, laser measurement and mechanical probing; and the underwater environment data is obtained through water quality sensors, optical imaging, magnetic detection and radar detection.

[0042] Preferably, the judgment method in step S4 includes judging the three dimensions of water depth, silt thickness and pollutants, wherein the silt removal threshold judgment based on water depth is determined by setting D 设计 is the designed water depth of the river, D 实际 is the actual measured water depth, and the water depth threshold coefficient for desilting is α. The desilting threshold judgment formula in the water depth dimension is:

[0043]

[0044] When it is greater than or equal to the threshold coefficient α, desilting is required;

[0045] The silt removal threshold is determined by setting B上限 is the maximum allowable silt thickness, B 实际 is the current measured silt thickness, and the silt thickness correction coefficient triggered by silt desilting is β, and the formula is:

[0046] B 实际 ≥B 上限 ×(1+β)

[0047] The desilting threshold based on pollutant content is determined by setting C 标准 is the environmental safety standard content of a pollutant, C 实际 is the actual pollutant content in the current sludge, and the pollutant content amplification factor triggered by dredging is γ, and the formula is:

[0048] C 实际 ≥C 标准 ×γ

[0049] By comprehensively judging whether the dredging threshold range has been reached, the dredging location and scale can be determined.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention uses a drone equipped with a laser radar combined with a bottom mud detection rod to quickly obtain underwater terrain information of a large area to be detected, construct a three-dimensional image, greatly improve the comprehensiveness and accuracy of terrain data acquisition, provide a reliable basis for subsequent analysis, accurately identify complex terrain locations, help focus on key areas, improve work pertinence, avoid blind detection in complex areas, effectively identify anomalies, discover potential geological risks in advance, judge key dredging locations and implement treatment, ensure the accurate implementation of dredging projects, improve dredging efficiency, reduce resource waste, maximize the benefits of dredging projects, effectively detect areas with complex underwater terrain, improve detection accuracy, and provide decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a flow chart of the geological exploration steps of the present invention. DETAILED DESCRIPTION

[0053] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are merely examples, and those skilled in the art may conceive of other obvious variations.

[0054] Reference Figure 1 As shown in the figure, the geological survey method for dredging in inland rivers and offshore water bodies includes the following steps:

[0055] S1. Use a drone equipped with a laser radar to fly over the area to be detected, and obtain underwater terrain information data of the area to be detected point by point through two-dimensional scanning. Use a bottom mud detection rod to detect depth data, and construct a three-dimensional image after pre-processing.

[0056] S2. Process the constructed 3D image elevation data based on edge detection methods, calculate the gradient of each pixel, set a gradient threshold to determine the edge points, connect the edge points to form a contour, and identify the complex terrain locations in the area to be detected;

[0057] S3. Extract data characteristic values ​​related to complex terrain locations from the three-dimensional model. Based on the preliminary terrain data reports of inland rivers and offshore water bodies obtained by the current geological bureau, construct a statistical model to calculate the average elevation range and area distribution interval statistical parameters of normal complex terrain areas. Compare the characteristic values ​​of each identified complex terrain location with the statistical parameters to identify whether there are any abnormalities.

[0058] S4. Use conventional methods to detect geological data at flat locations in the area to be detected. For the identified complex terrain locations, use underwater robots to detect and obtain underwater silt thickness data and underwater environment data. Based on the needs of the dredging project, establish a dredging threshold range, determine the key dredging locations, and carry out dredging treatment at the key dredging locations.

[0059] In this application, the drone is equipped with a lidar and can quickly cover a large area to be detected. It scans the underwater terrain point by point in two dimensions and detects depth data in conjunction with the bottom mud detection rod. The combination of the two greatly improves the comprehensiveness of data collection. After preprocessing, a three-dimensional image is constructed to intuitively present the overall picture of the underwater terrain, providing a rich and accurate data basis for subsequent analysis. Compared with a single detection method, it greatly improves the efficiency of terrain data acquisition; the three-dimensional image elevation data is processed by the edge detection method, the pixel gradient is calculated and the threshold is set to determine the edge point, thereby accurately outlining the contours of complex terrain. This method can keenly capture the details of terrain changes and effectively identify complex areas that are easily overlooked by traditional methods, which helps to focus subsequent work, improve the pertinence and accuracy of work, avoid blind detection in complex terrain areas, save manpower, material resources and time costs, extract characteristic values ​​of complex terrain data from the three-dimensional model, and construct a statistical model based on the preliminary terrain data report of the inland rivers and near-shore waters of the Geological Bureau to calculate the key statistical parameters of normal complex terrain areas. The method can provide a strong basis for potential geological risk assessment, help formulate response strategies in advance, reduce the risks of subsequent engineering construction, and ensure the safe and smooth progress of the project. Differentiated detection methods are adopted for different terrain characteristics, and conventional methods are used for flat locations, which is economical and efficient. Underwater robots are used for complex terrain. Their flexibility and high-precision sensors can obtain detailed underwater silt thickness and environmental data, meeting the high requirements of dredging projects for complex regional data, ensuring the accuracy and effectiveness of data acquisition, and providing reliable support for dredging project decision-making. Based on comprehensive and accurate data, the dredging threshold range is established, which can accurately determine the key dredging locations, avoid unnecessary dredging work, reduce resource waste, and carry out treatment on key dredging locations, so as to concentrate resources for efficient dredging, improve dredging efficiency, improve river water quality and ecological environment, maximize the comprehensive benefits of dredging projects, and achieve effective restoration and improvement of river functions.

[0060] In step S1, historical geological data is collected in advance to understand the geological background and plan the UAV route. During the flight, the lidar emits a laser beam at a high frequency, which is directed to the river surface and received. The UAV obtains underwater terrain data in a two-dimensional point-by-point scanning manner according to the preset route; the bottom mud detection rod is equipped with a pressure sensor to monitor the water pressure, and the depth of the detection rod is calculated based on the correspondence between water pressure and depth. Each preset point is detected more than three times; the acquired data is preprocessed, which includes data cleaning, data denoising and normalization; the preprocessed data is input into the modeling software to construct a three-dimensional model.

[0061] The advance collection of historical geological data is extremely critical for this application. The staff has a comprehensive and in-depth understanding of the geological background of the area to be explored by consulting geological archives, studying past geological exploration reports and communicating with local geological departments. These historical data cover many aspects of information such as stratigraphic structure, rock properties, and records of past geological changes, providing a solid knowledge reserve for the subsequent planning of drone routes. For example, if it is known that geological fault activity has occurred in a certain area, the detection points in that area can be specifically encrypted when planning the route to ensure that more detailed and accurate data are obtained.

[0062] When planning a UAV route, professional geographic information system software is used to comprehensively consider the shape, area, surrounding environment of the area to be detected, and key areas of focus suggested by historical geological data. Combined with the flight performance parameters of the UAV, such as cruising range, flight altitude limit, and scanning angle range, a scientific and reasonable route is designed to ensure that the UAV can fully cover the area to be detected and avoid collisions with obstacles during flight, while maximizing scanning efficiency. Route planning usually adopts grid or spiral scanning mode to ensure the uniformity and integrity of data collection; during the flight, the onboard laser radar plays a core role. The lidar equipment emits laser beams at an extremely high frequency, up to tens of thousands or even hundreds of thousands of times per second. These laser beams are shot at the river surface at a nearly vertical angle. When the laser propagates in the water, part of the energy will be absorbed and scattered by the water and the bottom of the water, while the remaining laser energy will be reflected back and received by the lidar. The drone flies stably in strict accordance with the preset route, and performs intensive data collection on the river surface in a two-dimensional point-by-point scanning manner. Each scanning point records the time difference between laser emission and reception. Through precise time measurement and the known speed of light, the distance from the point to the lidar can be calculated, and then the elevation information of the underwater terrain can be converted.

[0063] In step S2, a gradient calculation is performed for each pixel point in the 3D image elevation data. The calculation steps are as follows: first, the gradients of the pixel point in the horizontal and vertical directions are calculated, and then the gradients are combined to obtain the total gradient. Finally, the direction data of the gradient is calculated.

[0064] The three-dimensional image elevation data is a two-dimensional matrix Z(x,y), where x and y are the coordinates of the pixel point, and Z(x,y) represents the elevation value of the pixel point; calculate the horizontal gradient G x , calculated using the central difference formula, the expression is:

[0065]

[0066] Among them G x (x,y) means that in the x direction, the gradient of the current pixel is approximately half the difference between the elevation values ​​of the two adjacent pixels;

[0067] Calculate the vertical gradient G y , also using the central difference formula, the total gradient is obtained by combining the horizontal and vertical gradients, and the combined expression is:

[0068]

[0069] Among them, G(x,y) calculates the gradient size of each pixel point, reflecting the degree of change of the elevation value around the point. y (x, y) is in the y direction, and the gradient of the current pixel is approximately half of the difference in elevation between the two adjacent pixels. The direction of the gradient is calculated as θ, and the expression for the direction of the gradient is:

[0070]

[0071] Where θ(x,y) is the direction of the gradient at each pixel, in radians.

[0072] This application uses the central difference formula to calculate the gradient in the horizontal and vertical directions respectively, and can keenly perceive the subtle changes in the elevation values ​​around each pixel in different directions. In mountain rivers, the underwater terrain is undulating and changeable. This method can accurately identify even the smallest terrain differences or protrusions, providing delicate and accurate basic data for subsequent edge detection, thereby clearly outlining the contours of complex terrain and avoiding missing key terrain features. The horizontal and vertical gradients are combined to obtain the total gradient, which integrates the terrain change information in all directions in the two-dimensional plane. This makes the analysis of the terrain no longer limited to a single direction, and can comprehensively and three-dimensionally display the degree of change of the elevation values ​​around each pixel. Whether it is a gentle slope, a steep slope or irregular terrain, the total gradient value can accurately reflect the severity of its change, helping engineers quickly locate areas of terrain mutation and provide information for engineering decision-making. Strong basis; the calculated gradient direction can intuitively indicate the direction of terrain changes at each pixel point. When conducting geological analysis or dredging project planning, it is very important to understand the direction of terrain changes. For example, in river dredging, if the gradient direction of a certain area is known, it can be used to determine the main deposition direction of sediment carried by the water flow, thereby optimizing the dredging plan, improving dredging efficiency, and accurately determining the key dredging locations; the entire gradient calculation process is based on rigorous mathematical formulas, and each pixel point in the three-dimensional image elevation data is processed one by one to form a complete and systematic analysis method. Compared with simple threshold judgment or rough terrain analysis methods, this refined gradient calculation significantly improves the accuracy and reliability of complex terrain location identification, and can effectively distinguish different types of complex terrain, such as valleys, ridges, and faults, providing solid data support for subsequent targeted geological exploration and engineering operations.

[0073] In step S2, the gradient threshold is obtained based on experimental data. When the gradient value of the current pixel is greater than the threshold, the current pixel is determined to be an edge point. After the edge point is determined, the discrete edge points are connected to form a contour. Based on the edge tracking algorithm, pixels with gradient directions similar to the gradient direction of the current point and which are also edge points are found in the neighborhood of the point. The pixels are connected in sequence to form a closed contour. By analyzing the contour shape, area and direction characteristics, the location of complex terrain in the area to be detected is identified.

[0074] This application obtains gradient thresholds from experimental data and can accurately set them based on the actual terrain characteristics of the specific area to be detected. The geological conditions and terrain undulations in different areas vary greatly. Through preliminary field experiments in this area, a large amount of gradient data covering various terrain conditions is collected, and its distribution patterns are analyzed to determine the gradient threshold that is most suitable for this area. In this way, compared with the use of a universal fixed threshold, edge points can be more accurately determined, reducing edge misjudgments or omissions caused by unreasonable thresholds, and improving the accuracy of subsequent complex terrain recognition. After determining the edge points, an edge tracking algorithm is used to connect discrete points to form closed contours. This process can effectively integrate scattered terrain change information. In complex underwater environments, terrain changes are not continuous and regular. If discrete edge points are not reasonably connected, it is difficult to intuitively present the true shape of the terrain. By finding pixels with similar gradient directions and being edge points in the neighborhood and connecting them, a contour can be constructed based on the inherent logic of terrain changes, completely outlining the boundaries of complex terrain, and providing clear and coherent terrain contour information for subsequent analysis.

[0075] Based on the calculated pixel gradient value, the edge points of similar points are determined. When the following conditions are met, the pixel point is determined to be an edge point, namely:

[0076]

[0077] Edge(x,y) is a binary function, a value of 1 indicates that the pixel is an edge point, a value of 0 indicates that it is not an edge point, and T is the gradient threshold obtained;

[0078] Connect edge points to form contours by using the neighborhood N(x i ,y i ) to find other edge points (x j ,y j ), the connection is made when the following conditions are met: the neighborhood point is an edge point, that is, Edge(x j ,y j )=1; the gradient directions are similar. Let the threshold of similar gradient directions be Δθ, then it must satisfy:

[0079] |θ(x i y i)-θ(x j ,y j )|<Δθ

[0080] By continuously starting from an edge point, searching for edge points that meet the conditions in its neighborhood and connecting them, a closed contour is finally formed. For the closed contour formed Define the eigenvalues ​​to describe:

[0081] Contour area It is approximated by counting the number of pixels enclosed by the contour;

[0082] Profile shape factor This is measured using circularity, which is defined as: in is the perimeter of the contour;

[0083] Contour direction The main direction of the contour is determined by analyzing the coordinate change trend of each point on the contour and calculating it;

[0084] By analyzing the features, the location of complex terrain is identified; when the contour area is small, the shape factor deviates greatly from 1, and the direction is tortuous, the area surrounded by the contour is complex terrain, and a comprehensive judgment function is defined when The area enclosed by the contour is considered complex terrain when the following conditions are met:

[0085]

[0086] Here, ξ is a comprehensive judgment threshold, which is determined based on the specific application scenario and data characteristics.

[0087] This application uses a binary function combined with a gradient threshold T to determine whether a pixel point is an edge point. A value of 1 indicates that it is an edge point, and a value of 0 indicates that it is not. Within the neighborhood of the gradient edge point, the edge points are connected based on the conditions that the point is an edge point and the gradient direction is similar, and finally a closed contour is formed. The area is approximately calculated using the number of pixels surrounded by the contour; the shape factor is measured by circularity; the trend of changes in the coordinates of the contour points is analyzed to determine the direction and main direction; a comprehensive judgment function is constructed based on the contour area, shape factor, and direction characteristics. When specific conditions are met and compared with the comprehensive judgment threshold, it is determined whether the area surrounded by the contour is complex terrain. The edge points can be accurately determined to avoid misjudgment and missed judgment; the contours are reasonably connected to reflect the terrain boundary morphology; the contours are described in multiple dimensions to provide comprehensive information; complex terrain is efficiently identified, work efficiency is improved, and costs are reduced.

[0088] In step S3, the model is constructed by selecting polynomial regression as the model algorithm, and the constructed model is trained and optimized. The average elevation is calculated as:

[0089]

[0090] in is the average elevation, h i is the elevation value of the i-th sampling point, and n is the total number of sampling points;

[0091] The formula for calculating the elevation range is:

[0092] R=h max -h min

[0093] where h max is the maximum elevation in the area, h min is the minimum elevation value in the area;

[0094] The statistics of area distribution interval are calculated by using geometric formulas. If it is an irregular area, it is calculated based on the grid method. The area is divided into several small grids. The area of ​​each grid is known. The number of grids contained in the area is counted, and then the total area is multiplied by the area of ​​a single grid. The area distribution interval first determines the grouping interval of the area, and then counts the number of sub-areas falling in each interval, calculates the frequency of each interval, and obtains the statistical parameters of the area distribution interval.

[0095] Polynomial regression was chosen as the model algorithm for this application, as it effectively fits the nonlinear relationships between complex terrain data. Complex underwater terrain does not exhibit simple linear changes. Polynomial regression can flexibly adapt to terrain features of varying complexity by adjusting the degree of the polynomial, accurately modeling elevation data. This provides a reliable foundation for the subsequent calculation of average elevation and elevation range parameters, improving the model's ability to interpret actual terrain data and its prediction accuracy. The average elevation is calculated using a specific formula that comprehensively considers the elevation value of each sampling point. This means that the calculated result can fully reflect the overall elevation level within the area and avoid misjudgment of the overall elevation due to the influence of individual anomalous sampling points. In complex terrain with localized deep pits or protrusions, this calculation method can produce a more representative average elevation, providing accurate data for assessing the overall relief of the terrain. The elevation range is determined using a formula that clearly defines the difference between the maximum and minimum elevation values ​​within the area. This parameter intuitively reflects the relief of the terrain and is crucial for determining the complexity of the terrain. A larger elevation range often indicates a large drop in the terrain, potentially including steep slopes and complex canyons, helping to quickly identify areas of terrain that require special attention.

[0096] The comparison method in step S3 is based on the Euclidean distance method to perform comparison and judgment to identify whether there is complex terrain with abnormal judgment in the three-dimensional model. If there is a misjudgment of complex terrain, steps S1 and S2 are repeated to continue construction and identification, and abnormal judgment is performed again until there is no abnormality.

[0097] In order to accurately identify whether there is complex terrain with abnormal judgment in the three-dimensional model, this application uses the Euclidean distance method as the core comparison method. As a classic algorithm widely used in data analysis and pattern recognition, the Euclidean distance method can effectively measure the similarity between different data points. First, a set of key data points representing the characteristics of normal complex terrain areas is extracted from the three-dimensional model. These data points cover the average elevation range, area distribution range, and the values ​​corresponding to other relevant terrain feature parameters. For each complex terrain area to be judged, its corresponding feature data points are also extracted. Then, the distance between the feature points of the area to be judged and the feature points of the normal complex terrain area is calculated using the Euclidean distance formula.

[0098] Conventional methods in step S4 include geological drilling, geological radar and electrical exploration, which detect geological information in flat locations; detection is carried out based on underwater robots in complex terrain, and silt thickness data is obtained through sonar measurement, laser measurement and mechanical probing; underwater environment data is obtained through water quality sensors, optical imaging, magnetic detection and radar detection.

[0099] This application uses conventional methods such as geological drilling, geological radar and electrical exploration in flat locations to efficiently obtain geological information. Geological drilling can directly collect core samples, intuitively understand the stratum structure and rock and soil characteristics, and provide accurate deep geological data for the project. Geological radar uses electromagnetic waves to detect the distribution of underground media, quickly draw stratum images, clearly present different stratum interfaces and potential anomalies, and help quickly locate geological change areas. Electrical exploration detects underground geological structures based on the differences in electrical properties of different geological bodies, and has a significant effect on finding aquifers and faults. It fully reveals the geological conditions of flat areas and provides a solid basis for subsequent engineering planning. For complex terrain, underwater robots are equipped with a variety of equipment for detection. When obtaining silt thickness data, sonar measurements calculate silt thickness by transmitting and receiving sound waves and scanning a large area. It has a wide coverage and It is highly efficient and has extremely high laser measurement accuracy. When the water is clear, it can accurately measure the distance to the silt surface, providing reliable data for accurately grasping the silt thickness. Mechanical probing directly senses the physical properties of the silt and accurately judges the thickness. The three methods complement each other to ensure that the silt thickness data under complex terrain is comprehensive and accurate. In terms of obtaining underwater environmental data, water quality sensors monitor water temperature, salinity, and pH parameters in real time to quickly grasp the dynamic changes in water quality. Optical imaging provides intuitive underwater images, clearly showing the landform and biological conditions, and assisting in understanding the ecological environment. Magnetic detection detects metal objects and changes in geological structures. Radar detection penetrates the underwater medium to reveal the stratigraphic structure and anomalies. Combining multiple means, underwater environmental data is obtained in all directions, helping to comprehensively evaluate the underwater conditions under complex terrain, provide sufficient information for dredging projects and ecological protection, and ensure the scientific and reasonable implementation of the project.

[0100] The judgment method in step S4 includes judging the three dimensions of water depth, silt thickness and pollutants, wherein the silt removal threshold judgment based on water depth is determined by setting D 设计 is the designed water depth of the river, D 实际 is the actual measured water depth, and the water depth threshold coefficient for desilting is α. The desilting threshold judgment formula in the water depth dimension is:

[0101]

[0102] When it is greater than or equal to the threshold coefficient α, desilting is required;

[0103] The silt removal threshold is determined by setting B 上限 is the maximum allowable silt thickness, B 实际 is the current measured silt thickness, and the silt thickness correction coefficient triggered by silt desilting is β, and the formula is:

[0104] B 实际 ≥B 上限 ×(1+β)

[0105] The desilting threshold based on pollutant content is determined by setting C标准 is the environmental safety standard content of a pollutant, C 实际 is the actual pollutant content in the current sludge, and the pollutant content amplification factor triggered by dredging is γ, and the formula is:

[0106] C 实际 ≥C 标准 ×γ

[0107] By comprehensively judging whether the dredging threshold range has been reached, the dredging location and scale can be determined.

[0108] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. Various changes and improvements are possible without departing from the spirit and scope of the present invention, and such changes and improvements fall within the scope of the invention as claimed.

Claims

1. A geological exploration method for dredging in inland rivers and offshore water bodies, characterized in that: The steps of geological exploration are: S1. Use a drone equipped with a laser radar to fly over the area to be detected, and obtain underwater terrain information data of the area to be detected by two-dimensional point-by-point scanning; The depth data is detected by the bottom mud detection rod and a three-dimensional image is constructed after pre-processing; S2. Process the constructed 3D image elevation data based on edge detection methods, calculate the gradient of each pixel, set a gradient threshold to determine the edge points, connect the edge points to form a contour, and identify the complex terrain locations in the area to be detected; S3. Extract data characteristic values ​​related to complex terrain locations from the three-dimensional model. Based on the preliminary terrain data reports of inland rivers and offshore water bodies obtained by the current geological bureau, construct a statistical model to calculate the average elevation range and area distribution interval statistical parameters of normal complex terrain areas. Compare the characteristic values ​​of each identified complex terrain location with the statistical parameters to identify whether there are any abnormalities. S4. Use conventional methods to detect geological data at flat locations in the area to be detected. For the identified complex terrain locations, use underwater robots to detect and obtain underwater silt thickness data and underwater environment data. Based on the needs of the dredging project, establish a dredging threshold range, determine the key dredging locations, and carry out dredging treatment at the key dredging locations.

2. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: In step S1, historical geological data is collected in advance to understand the geological background and plan the UAV route. During the flight, the lidar emits a laser beam at a high frequency, which is directed to the river surface and received. The UAV follows the preset route and obtains underwater terrain data in a two-dimensional point-by-point scanning manner; The bottom mud detection rod is equipped with a pressure sensor to monitor the water pressure. Based on the corresponding relationship between water pressure and depth, the depth of the detection rod is calculated. Each preset point is detected more than three times. Preprocess the acquired data, including data cleaning, data denoising and normalization; The preprocessed data is input into the modeling software to construct a three-dimensional model.

3. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: In step S2, a gradient calculation is performed for each pixel point in the 3D image elevation data. The calculation steps are as follows: first, the gradients of the pixel point in the horizontal and vertical directions are calculated, and then the gradients are combined to obtain the total gradient. Finally, the direction data of the gradient is calculated. The three-dimensional image elevation data is a two-dimensional matrix Z(x,y), where x and y are the coordinates of the pixel point, and Z(x,y) represents the elevation value of the pixel point; calculate the horizontal gradient G x , calculated using the central difference formula, the expression is: Among them G x (x,y) means that in the x direction, the gradient of the current pixel is approximately half the difference between the elevation values ​​of the two adjacent pixels; Calculate the vertical gradient G y , also using the central difference formula, the total gradient is obtained by combining the horizontal and vertical gradients, and the combined expression is: Among them, G(x,y) calculates the gradient size of each pixel point, reflecting the degree of change of the elevation value around the point. y (x, y) is in the y direction, and the gradient of the current pixel is approximately half of the difference in elevation between the two adjacent pixels. The direction of the gradient is calculated as θ, and the expression for the direction of the gradient is: Where θ(x,y) is the direction of the gradient at each pixel, in radians.

4. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: In step S2, the gradient threshold is obtained based on experimental data. When the gradient value of the current pixel is greater than the threshold, the current pixel is determined to be an edge point. After the edge point is determined, the discrete edge points are connected to form a contour. Based on the edge tracking algorithm, pixels with gradient directions similar to the gradient direction of the current point and which are also edge points are found in the neighborhood of the point. The pixels are connected in sequence to form a closed contour. By analyzing the contour shape, area and direction characteristics, the location of complex terrain in the area to be detected is identified.

5. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 4 is characterized in that: Based on the calculated pixel gradient value, the edge points of similar points are determined. When the following conditions are met, the pixel point is determined to be an edge point, namely: Edge(x,y) is a binary function. A value of 1 indicates that the pixel is an edge point, and a value of 0 indicates that it is not an edge point. T is the gradient threshold obtained. Connect edge points to form contours by using the neighborhood N(x i ,y i ) to find other edge points (x j ,y j ), the connection is made when the following conditions are met: the neighborhood point is an edge point, that is, Edge(x j ,y j )=1; the gradient directions are similar. Let the threshold of similar gradient directions be Δθ, then it must satisfy: |θ(x i y i )-θ(x j ,y j )|<Δθ By continuously starting from an edge point, searching for edge points that meet the conditions in its neighborhood and connecting them, a closed contour is finally formed. For the closed contour formed Define the eigenvalues ​​to describe: Contour area It is approximated by counting the number of pixels enclosed by the contour; Profile shape factor This is measured using circularity, which is defined as: in is the perimeter of the contour; Contour direction The main direction of the contour is determined by analyzing the coordinate change trend of each point on the contour and calculating it; By analyzing the features, the location of complex terrain is identified; when the contour area is small, the shape factor deviates greatly from 1, and the direction is tortuous, the area surrounded by the contour is complex terrain, and a comprehensive judgment function is defined when The area enclosed by the contour is considered complex terrain when the following conditions are met: Here, ξ is a comprehensive judgment threshold, which is determined based on the specific application scenario and data characteristics.

6. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: In step S3, the model is constructed by selecting polynomial regression as the model algorithm, and the constructed model is trained and optimized. The average elevation is calculated as: in is the average elevation, h i is the elevation value of the i-th sampling point, and n is the total number of sampling points; The formula for calculating the elevation range is: R=h max -h min where h max is the maximum elevation in the area, h min is the minimum elevation value in the area; The statistics of area distribution interval are calculated by using geometric formulas. If it is an irregular area, it is calculated based on the grid method. The area is divided into several small grids. The area of ​​each grid is known. The number of grids contained in the area is counted, and then the total area is multiplied by the area of ​​a single grid. The area distribution interval first determines the grouping interval of the area, and then counts the number of sub-areas falling in each interval, calculates the frequency of each interval, and obtains the statistical parameters of the area distribution interval.

7. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: The comparison method in step S3 is based on the Euclidean distance method to perform comparison and judgment to identify whether there is complex terrain with abnormal judgment in the three-dimensional model. If there is a misjudgment of complex terrain, steps S1 and S2 are repeated to continue construction and identification, and abnormal judgment is performed again until there is no abnormality.

8. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1 is characterized in that: Conventional methods in step S4 include geological drilling, geological radar and electrical exploration, which detect geological information in flat locations; detection is carried out based on underwater robots in complex terrain, and silt thickness data is obtained through sonar measurement, laser measurement and mechanical probing; underwater environment data is obtained through water quality sensors, optical imaging, magnetic detection and radar detection.

9. The dredging geological exploration method for inland rivers and offshore water bodies according to claim 1, characterized in that: The judgment method in step S4 includes judging the three dimensions of water depth, silt thickness and pollutants, wherein the silt removal threshold judgment based on water depth is determined by setting D 设计 is the designed water depth of the river, D 实际 is the actual measured water depth, and the water depth threshold coefficient for desilting is α. The desilting threshold judgment formula in the water depth dimension is: When it is greater than or equal to the threshold coefficient α, desilting is required; The silt removal threshold is determined by setting B 上限 is the maximum allowable silt thickness, B 实际 is the current measured silt thickness, and the silt thickness correction coefficient triggered by silt desilting is β, and the formula is: B 实际 ≥B 上限 ×(1+β) The desilting threshold based on pollutant content is determined by setting C 标准 is the environmental safety standard content of a pollutant, C 实际 is the actual pollutant content in the current sludge, and the pollutant content amplification factor triggered by dredging is γ, and the formula is: C 实际 ≥C 标准 ×γ By comprehensively judging whether the dredging threshold range has been reached, the dredging location and scale can be determined.

Images