Earthwork balance calculation method, system, device and storage medium

Abstract

The present application relates to a kind of earthwork balance calculation method, system, equipment and storage medium, it is related to earthwork balance design technical field.The method includes fusion processing to oblique photography data and laser radar point cloud data, generates real scene three-dimensional model;Real scene three-dimensional model is carried out grid division, and grid model is obtained;Based on grid model, definition includes the design elevation of each grid point, site overall slope, the slope height of each side slope unit and the decision variable set of earthwork allocation amount, and construct integrated objective function including earthwork balance target, economy target and safety target;Under engineering constraint condition, to integrated objective function is carried out multi-objective optimization solution, obtains Pareto optimal solution set, and from Pareto optimal solution set selected final implementation scheme.The present application effectively solves the problem of terrain data missing in dense vegetation area, breaks through the limitation of traditional single-objective optimization, realizes the maximization of comprehensive benefit.

Description

Technical Field

[0001] This invention relates to the field of earthwork balance design technology, and in particular to an earthwork balance calculation method, system, equipment and storage medium. Background Technology

[0002] In power grid construction, earthwork is a fundamental aspect of substation and transmission line construction, and its design quality directly impacts project investment, construction period, operational safety, and environmental friendliness. Earthwork balance design, as a core component of earthwork engineering, aims to achieve a near-perfect match between excavation and filling volumes on-site through precise calculation and optimized allocation. This minimizes the need for borrowed and discarded soil, thereby effectively controlling project costs, shortening the construction period, and reducing environmental impact.

[0003] Currently, the most commonly used earthwork balance design methods in the industry mainly rely on designers using measurement data and personal experience to calculate and allocate earthwork volumes using grid methods or digital terrain modeling (DTM) methods (such as the irregular triangular network method). While these methods have played a role in past practice, their technical limitations are becoming increasingly apparent as the scale and complexity of power grid construction continue to expand, primarily in the two key areas of high-precision terrain data acquisition and earthwork balance calculation.

[0004] I. Topographic Data Acquisition

[0005] Traditional topographic surveying methods mainly include manual surveying, single-UAV oblique photogrammetry, and 3D laser scanning. Manual surveying typically uses equipment such as total stations and GPS-RTK for single-point data acquisition. This method is inefficient, time-consuming, and can only acquire discrete point elevation data, making it difficult to create high-precision continuous terrain models. It is particularly prone to data omissions in complex terrain areas, failing to meet the requirements of refined design. While single-UAV oblique photogrammetry offers advantages such as high efficiency and low cost, and can quickly generate realistic 3D models from multi-angle images, its data quality is significantly affected by factors such as vegetation cover and weather conditions. In densely vegetated areas, under building eaves, or in areas with abrupt terrain changes, problems such as model voids, inconsistencies, and texture distortion can easily occur, making it difficult to obtain accurate ground elevations and guaranteeing precision. Although 3D laser scanning can acquire high-precision point cloud data, it suffers from problems such as large data volume, high point cloud dispersion, complex subsequent processing, and poor visualization of optimized design schemes. The methods described above are inadequate in terms of data accuracy, completeness, and reliability, making it difficult to support the high-quality input data requirements of subsequent multi-objective optimization algorithms.

[0006] II. Earthwork Balance Calculation Methods

[0007] Traditional earthwork calculations primarily employ the grid method and the DTM method. The grid method divides the site into several regular grids (such as square grids), calculates cut and fill volumes grid by grid and sums them up, making it suitable for flat areas, but its accuracy is limited in complex terrain. The DTM method constructs a triangular network (TIN) model based on terrain feature points, which better adapts to terrain variations, but still heavily relies on the designer's experience and manual intervention. Traditional earthwork balance calculation methods heavily depend on the designer's experience and have shortcomings in rapid comparison of multiple options and global optimization.

[0008] ① High dependence on human experience: Key decisions such as determining the initial design elevation, adjusting the direction and extent, and earthwork allocation plan all depend on the engineer's personal experience, lacking a systematic and quantitative optimization mechanism;

[0009] ② The iterative process is cumbersome and inefficient: the pursuit of "balance" and "economy" requires repeated manual adjustments to the design elevation and recalculation, which involves a large amount of calculation work and a long cycle;

[0010] ③ Difficult to achieve global optimum: Traditional methods are essentially a "semi-manual" trial and error process, making it difficult to quickly find the optimal solution with the lowest total cost or the smallest earthwork volume from a massive number of possible solutions;

[0011] ④ Limited factors to consider: Traditional calculations mainly focus on earthwork volume balance, making it difficult to simultaneously incorporate multiple objectives such as transportation costs, slope safety, and environmental impact into quantitative considerations and comprehensive optimization.

[0012] Based on the current state of technology and engineering practice, the following obvious limitations have been identified in traditional methods:

[0013] (1) Low accuracy and low efficiency: It relies on manual collection and manual calculation, which results in slow data collection speed, low processing efficiency, and easy introduction of human error, leading to large deviations in earthwork volume calculation;

[0014] (2) Single optimization objective and lack of global coordination: If only "minimum earthwork volume" or "fill-cut balance" is taken as a single objective, it is difficult to comprehensively consider the coordinated optimization of multiple objectives such as economy, safety and environmental protection. For example, the design often does not fully consider slope stability, earthwork transportation costs and the impact on the surrounding ecological environment.

[0015] (3) Difficulty in selecting schemes and high dependence on experience: For complex sites, designers can only generate a limited number of schemes for comparison and selection. They cannot automatically generate and evaluate a large number of feasible schemes and quickly locate the optimal solution. The design quality is largely dependent on personal experience.

[0016] (4) Poor visualization and interactivity: The design results are mostly presented in the form of two-dimensional drawings and digital tables, lacking intuitive three-dimensional visualization, which is not conducive to design review and multi-party collaboration, and it is also difficult to predict the effect of the combination of the design scheme and the actual terrain.

[0017] (5) Difficult to cope with complex terrain and special geological conditions: In special geological conditions such as mountainous areas and soft soil layers, the solutions provided by traditional methods are often not targeted enough, which can easily lead to increased investment or safety hazards.

[0018] Therefore, there is an urgent need for an earthwork balance calculation method that can integrate high-precision terrain data acquisition, multi-objective collaborative optimization, and three-dimensional visualization to improve the scientific, economic, and environmentally friendly nature of earthwork design for power grid projects. Summary of the Invention

[0019] Therefore, it is necessary to provide an earthwork balance calculation method, system, equipment, and storage medium to address at least one of the following problems of traditional methods: low accuracy and efficiency, single optimization objective and lack of global coordination, difficulty in scheme comparison and selection, high dependence on experience, poor visualization and interactivity, and difficulty in dealing with complex terrain and special geological conditions.

[0020] In a first aspect, the present invention provides a method for earthwork balance calculation, comprising:

[0021] Acquire oblique photography data and lidar point cloud data of the target area;

[0022] The oblique photography data and lidar point cloud data are fused to generate a high-precision real-scene 3D model.

[0023] The real-world 3D model is divided into meshes to obtain a meshed model for optimization calculations;

[0024] Based on the gridded model, a set of decision variables is defined, including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit, and the earthwork allocation. A comprehensive objective function containing earthwork balance, economic and safety objectives is constructed based on the set of decision variables.

[0025] Under the preset engineering constraints, the comprehensive objective function is solved by multi-objective optimization to obtain the Pareto optimal solution set, and the final implementation scheme is selected from the Pareto optimal solution set.

[0026] In this invention, oblique photogrammetry data can provide rich textures and continuous models, while lidar penetrating vegetation can provide accurate elevations. The fused data combines the advantages of both, effectively solving the problem of missing terrain data in densely vegetated areas. It achieves centimeter-level accuracy in obtaining real surface elevations across the entire region, laying a reliable data foundation for subsequent accurate earthwork volume calculations.

[0027] By simultaneously incorporating economic efficiency (cost), safety (slope stability), and earthwork balance into the automated optimization process, this multi-objective optimization can automatically find the best solution from a massive number of options, accurately recommending the solution with the minimum total earthwork volume, the lowest earthwork transportation cost, and the most economical slope support. It outputs a Pareto optimal solution set that achieves the best balance among multiple objectives, providing decision-makers with a rich and scientific selection of optimal solutions. This breaks through the limitations of traditional single-objective optimization and maximizes comprehensive benefits.

[0028] Compared with traditional trial-and-error methods that rely on experience, it can reduce the direct cost of earthwork engineering by 10% to 20%. At the same time, by optimizing slope design, it can significantly improve the slope safety factor and reduce the operation and maintenance risks and disposal costs of geological disasters such as landslides and collapses in the later stage.

[0029] This invention achieves full-process automation. Designers only need to set optimization goals and constraints to automatically complete tens of thousands of calculations and scheme comparisons, improving the efficiency of scheme generation and optimization by 5 to 10 times, allowing designers to focus on higher-value decision-making work.

[0030] This invention provides intuitive, visual decision support, improving review and decision-making efficiency. The final design outcome is no longer an abstract two-dimensional drawing or numerical table, but a realistic three-dimensional model fully integrated with the real environment. Decision-makers can intuitively review the cut and fill areas, slope morphology, and road orientation of different schemes in a three-dimensional scene, and conduct interactive cross-sectional analysis, slope analysis, and line-of-sight analysis, greatly improving the efficiency of multi-party collaborative review and the understandability of the scheme.

[0031] In one embodiment, acquiring oblique photography data and lidar point cloud data of the target area includes:

[0032] The target area is surveyed, and based on the terrain complexity and vegetation coverage, a drone oblique photography flight plan and a lidar scanning strategy are planned in a coordinated manner.

[0033] The UAV oblique photography flight scheme includes a grid-shaped flight path or a five-lens camera acquisition scheme, ensuring that the overlap between the flight path and the lateral direction is not less than 60% to obtain ground texture information and macroscopic geometric shape; the lidar scanning strategy includes selecting at least one of airborne, ground-based, or handheld SLAM methods for scanning according to data gap filling requirements, and setting the scanning distance, scanning density, and scanning angle to actively penetrate vegetation gaps and obtain high-precision three-dimensional point clouds of the ground surface and hidden areas.

[0034] Based on the UAV oblique photography flight scheme, the oblique photography data is acquired; based on the lidar scanning strategy, the lidar point cloud data is acquired.

[0035] In one embodiment, the oblique photography data and lidar point cloud data are fused, including:

[0036] Aerial triangulation and dense matching are performed on the oblique photogrammetry data to generate densely matched image point clouds;

[0037] The lidar point cloud data is denoised, filtered, and classified to extract clean ground point clouds;

[0038] The image point cloud and the ground point cloud are coarsely and finely registered to generate fused point cloud data.

[0039] Based on the fused point cloud data, a high-precision real-scene 3D model is constructed.

[0040] In this invention, lidar can effectively penetrate vegetation gaps to directly acquire true surface elevation point cloud data of obscured areas. This complements the rich texture information from oblique photogrammetry, solving the technical challenges of data loss, model voids, and distortion caused by single data sources in complex terrains such as densely vegetated areas, steep slopes, and building bases. This achieves flawless data acquisition across the entire domain. Coarse registration adjusts the position and orientation of the image point cloud and the ground point cloud to a more similar state, improving the success rate of fine registration and reducing the number of iterations required. Fine registration eliminates residual minor deviations, achieving high-precision registration.

[0041] In one embodiment, a modified iterative nearest-point algorithm is used for fine registration, including:

[0042] The ground point cloud after coarse registration is used as the target point cloud, and the image point cloud is used as the source point cloud to be registered.

[0043] For each point in the source point cloud, find its closest Euclidean distance counterpart in the target point cloud to form the current set of matched point pairs;

[0044] Based on the currently matched set of point pairs, an objective function is constructed, and the rigid body transformation matrix that minimizes the sum of the distances between corresponding points is solved; wherein, the rigid body transformation matrix includes a rotation matrix and a translation vector, and the initial value of the rigid body transformation matrix is ​​the transformation matrix obtained by coarse registration;

[0045] Based on prior knowledge or coarse registration results, determine the reasonable range of rotation angle variation in the current iteration;

[0046] The rotation matrix is ​​converted into Euler angles, and the Euler angles are constrained based on the reasonable range of variation of the rotation angles in the current iteration.

[0047] Convert the constrained Euler angles into rotation matrices;

[0048] A new rigid body transformation matrix is ​​formed by the constrained rotation matrix and the translation vector;

[0049] Update the source point cloud based on the new rigid body transformation matrix;

[0050] Determine whether the dynamic convergence condition has been met. If yes, output the final rigid body transformation matrix; otherwise, proceed to the step of forming the current set of matched point pairs and perform the next iteration.

[0051] In this invention, the introduction of rotation angle constraints and dynamic convergence conditions significantly improves the registration accuracy and efficiency of data from different sources, and stably controls the plane and elevation errors at the centimeter level, laying a reliable foundation for subsequent accurate calculations.

[0052] In one embodiment, the earthwork balance target is:

[0053] ;

[0054] in, Indicates the earthwork balance target; This represents the excavation volume of the k-th grid cell; This represents the fill volume of the k-th grid cell; N represents the number of grid cells. Indicates the penalty coefficient; and These represent the total excavation volume and the total fill volume, respectively. , ;

[0055] The economic objective is as follows:

[0056] ;

[0057] in, Indicates economic objectives; This represents the cost coefficient per unit of earthwork excavation or filling. This represents the unit distance transportation cost coefficient for a unit volume of earthwork. This represents the transportation distance from the i-th cut unit to the j-th fill unit; This represents the amount of earthwork allocated from the i-th excavation unit to the j-th fill unit; This represents the cost coefficient for slope protection per unit area. The total surface area of ​​the slope is represented by m; the number of cut units is represented by n; and the number of fill units is represented by n.

[0058] The security objective is:

[0059] or ;

[0060] in, Indicates security objectives; This indicates the safety factor of the slope.

[0061] This invention is the first to simultaneously optimize "lowest economic cost", "optimal earthwork balance" and "highest engineering safety" as parallel optimization objectives in power grid earthwork balance design, overcoming the limitations of traditional methods that only focus on a single earthwork volume objective, and maximizing comprehensive benefits.

[0062] In one embodiment, the engineering constraints include geometric constraints, elevation boundary constraints, earthwork balance constraints, and safety constraints;

[0063] The geometric constraint is that the slope between any two adjacent grid points is less than or equal to the maximum allowable slope; the elevation boundary constraint is that the design elevation of any grid point is between the minimum and maximum allowable design elevations; the earthwork balance constraint is:

[0064] ;

[0065] in, Indicates the total excavation volume; Indicates the total fill volume; Indicates the final looseness coefficient; Indicates the initial looseness coefficient; Indicates the volume of excavated soil; Indicates the amount of earthwork borrowed;

[0066] The safety constraint is that the slope safety factor is greater than the safety factor required by the specifications or technical requirements, and the slope height of any slope unit shall not exceed its maximum allowable height.

[0067] In this invention, engineering constraints not only ensure that each generated solution conforms to basic physical laws, technical specifications, and safety requirements, but also guide the algorithm to directly find the optimal solution within a meaningful solution space, avoiding optimization on invalid or erroneous solutions and improving optimization efficiency. Specifically, geometric constraints ensure that the site slope meets drainage and access requirements, avoiding unrealistic steep slopes; elevation boundary constraints ensure the site design is consistent with the surrounding environment and overall planning; earthwork balance constraints ensure that the optimization results achieve practical earthwork balance while considering the true physical properties (looseness) of the soil, fundamentally avoiding calculation errors in engineering quantities; and safety constraints embed slope stability design into the optimization process, eliminating unsafe solutions at the source and solving the problems of limited consideration factors and safety hazards.

[0068] In one embodiment, a non-dominated sorting genetic algorithm is used to solve the comprehensive objective function for multi-objective optimization, including:

[0069] The set of decision variables is encoded into chromosomes using real numbers, and an initial population is generated.

[0070] Calculate the function value of each individual in the population corresponding to the earthwork balance target, economic target and safety target, and verify its satisfaction with the engineering constraints;

[0071] Perform a fast nondominated ordination on each individual in the population to determine its nondominated rank, and calculate the crowding distance for individuals within the same nondominated rank.

[0072] Based on the non-dominant rank and crowding distance, offspring populations are generated through selection, crossover, and mutation operations;

[0073] The parent and offspring populations are merged, and a new generation of population is selected from the merged populations based on non-dominated ranking and crowding distance calculations.

[0074] Determine if the termination condition is met. If yes, output the Pareto optimal solution set composed of the individuals with the highest non-dominant level in the final population. If no, proceed to the step of calculating the function value of each individual in the population corresponding to the earthwork balance target, economic target, and safety target, and perform the next iteration loop.

[0075] In this invention, through a collaborative mechanism of adaptive crossover mutation operator, elite retention strategy, and crowding distance calculation, the algorithm can efficiently and uniformly explore the entire Pareto front, thereby generating a widely covered non-dominated solution set. This provides decision-makers with a series of optimal solutions that achieve the best trade-off among multiple competing objectives. Furthermore, the optimization algorithm deeply integrates key engineering constraints such as slope stability and soil looseness coefficient, fundamentally ensuring that all output solutions possess mathematical optimality, engineering feasibility, and safety reliability.

[0076] Secondly, the present invention also provides an earthwork balance calculation system, comprising:

[0077] The acquisition unit is used to acquire oblique photography data and lidar point cloud data of the target area;

[0078] The fusion processing unit is used to fuse the oblique photography data and the lidar point cloud data to generate a high-precision real-scene 3D model.

[0079] Mesh division unit, used to divide the real scene 3D model into meshes to obtain a meshed model for optimization calculation;

[0080] Define and construct a unit, which is used to define a set of decision variables based on the gridded model, including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit and the earthwork allocation, and construct a comprehensive objective function including earthwork balance objective, economic objective and safety objective based on the set of decision variables;

[0081] The solution and selection unit is used to perform multi-objective optimization on the comprehensive objective function under preset engineering constraints, obtain the Pareto optimal solution set, and select the final implementation scheme from the Pareto optimal solution set.

[0082] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instructions stored in the memory, wherein the processor executes the computer program / instructions to implement the steps in the earthwork balance calculation method described above.

[0083] Fourthly, the present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps in the earthwork balance calculation method described above.

[0084] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0085] This invention automates the entire process from data acquisition, processing, optimization to solution output, enabling the automatic batch generation and comparison of thousands of solutions. Compared to traditional methods that rely on manual trial and error and take several days, this invention shortens the design cycle to a few hours, improving design efficiency by 5-10 times. The optimization results can be directly output as BIM models or GIS standard formats, seamlessly connecting to subsequent refined design, construction layout, and digital operation and maintenance management platforms. This achieves full-process digital flow from design to construction, enhancing the management capabilities throughout the entire lifecycle. The final design results are presented in an immersive, high-precision, realistic 3D model, allowing designers to perform arbitrary sectioning, navigation, and quantity lookup, greatly improving review efficiency and the comprehensibility of the solutions. Attached Figure Description

[0086] Figure 1 Here is a flowchart of the earthwork balance calculation method;

[0087] Figure 2 A flowchart for fusing oblique photogrammetry data and lidar point cloud data;

[0088] Figure 3 This is a flowchart of a multi-objective optimization solution using a non-dominated sorting genetic algorithm;

[0089] Figure 4 This is a block diagram of the earthwork balance calculation system.

[0090] Figure 5 This is a schematic diagram of the electronic device structure. Detailed Implementation

[0091] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other.

[0092] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0093] The following describes some embodiments of the baby bottle according to the present invention with reference to the accompanying drawings.

[0094] Example 1

[0095] like Figure 1 As shown, the earthwork balance calculation method disclosed in this embodiment includes the following steps:

[0096] Step 1: Acquire oblique photography data and LiDAR point cloud data of the target area;

[0097] Step 2: The oblique photogrammetry data and lidar point cloud data obtained in Step 1 are fused to generate a high-precision real-world 3D model.

[0098] Step 3: Mesh the real-world 3D model generated in Step 2 to obtain a meshed model for optimization calculation;

[0099] Step 4: Based on the gridded model obtained in Step 3, define a set of decision variables including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit, and the earthwork allocation amount, and construct a comprehensive objective function that includes earthwork balance objective, economic objective, and safety objective based on the set of decision variables.

[0100] Step 5: Under the preset engineering constraints, perform multi-objective optimization on the comprehensive objective function constructed in Step 4 to obtain the Pareto optimal solution set, and select the final implementation scheme from the Pareto optimal solution set.

[0101] In this embodiment, the lidar can effectively penetrate the gaps in vegetation and directly acquire the real surface elevation point cloud data of the obscured area. It complements the rich texture information of the oblique photogrammetry data, solving the technical problem of data loss, model voids and distortion in complex terrains such as densely vegetated areas, steep slopes and the bottom of buildings by a single data source, and realizing data acquisition without any loopholes in the whole area.

[0102] This embodiment is the first to simultaneously optimize "lowest economic cost", "optimal earthwork balance" and "highest engineering safety" as parallel optimization objectives in power grid earthwork balance design, overcoming the limitations of traditional methods that only focus on a single earthwork volume objective, and maximizing comprehensive benefits.

[0103] The calculation method provided in this embodiment automates the entire process from data acquisition, processing, optimization to solution output. It can automatically generate and compare tens of thousands of solutions in batches, reducing the traditional design cycle, which relies on manual trial and error and takes several days, to a few hours, improving design efficiency by 5-10 times. The optimization results can be directly output as BIM models or GIS standard formats, seamlessly connecting to subsequent refined design, construction layout, and digital operation and maintenance management platforms. This achieves full digital flow from design to construction, enhancing the management capabilities throughout the entire lifecycle. The final design result is presented in an immersive form as a high-precision, realistic 3D model, allowing designers to perform arbitrary sectioning, navigation, and quantity lookup, greatly improving review efficiency and the comprehensibility of the solution.

[0104] In addition to the features of the above embodiments, this embodiment further defines the steps for acquiring oblique photography data and lidar point cloud data of the target area, specifically including:

[0105] Step 1.1: Conduct a survey of the target area, and based on the terrain complexity and vegetation coverage, collaboratively plan the UAV oblique photography flight scheme and the lidar scanning strategy.

[0106] Before the operation, the target area was first surveyed. Based on the terrain complexity and vegetation coverage, a UAV oblique photogrammetry flight plan and a LiDAR scanning strategy were designed. The UAV oblique photogrammetry flight plan is mainly responsible for acquiring rich texture information and macroscopic geometric shapes of ground features over a large area. In the design, a grid-like flight path or a five-lens camera acquisition scheme can be selected, ensuring that the overlap between the flight path and the lateral direction is not less than 60% to ensure image matching accuracy. In the LiDAR scanning strategy, different scanning strategies such as airborne, ground-based, or handheld SLAM are selected according to the data gap filling requirements, and parameters such as scanning distance, scanning density, and scanning angle are set. The LiDAR actively emits laser pulses, which can effectively penetrate the gaps in vegetation and directly acquire high-precision 3D point clouds of the ground surface (accuracy up to centimeter level), perfectly compensating for the data loss in oblique photogrammetry under vegetation, at the bottom of buildings, and in shaded areas.

[0107] During field data collection, a certain number of ground control points need to be deployed across the entire survey area, evenly distributed. These ground control points are ground markers with precisely known three-dimensional coordinates (i.e., true coordinates), which can be clearly identified in oblique photogrammetry data. When processing the oblique photogrammetry data, the coordinates of these ground control points are measured on an aerial triangulation model and then compared with their true coordinates to correct for GPS positioning errors and IMU attitude errors in the UAV.

[0108] Step 1.2: Obtain oblique photography data based on the UAV oblique photography flight scheme; obtain lidar point cloud data based on the lidar scanning strategy.

[0109] like Figure 2 As shown, in addition to the features of the above embodiments, this embodiment further defines the steps for fusing oblique photogrammetry data and lidar point cloud data, specifically including:

[0110] Step 2.1: Perform aerial triangulation and dense matching on the oblique photogrammetry data, solve the high-precision exterior orientation elements of each image, and generate densely matched image point clouds;

[0111] Step 2.2: Denoise, filter, and classify the lidar point cloud data (e.g., separate ground points, vegetation points, and building points) to extract clean ground point clouds;

[0112] Step 2.3: Perform coarse and fine registration on the image point cloud and the ground point cloud to generate fused point cloud data;

[0113] Step 2.4: Construct a high-precision real-scene 3D model based on the fused point cloud data.

[0114] Registration is crucial for multi-source data fusion. Before registration, the image point cloud and ground point cloud are unified to the CGCS2000 coordinate system or the project design coordinate system. In this embodiment, ground control points are first used to perform coarse registration of the image point cloud and ground point cloud, and then an improved Iterative Closest Point (ICP) algorithm is used for fine registration. Coarse registration adjusts the position and orientation of the image point cloud and ground point cloud to a more similar state, improving the success rate of fine registration and reducing the number of iterations required for fine registration; fine registration can eliminate residual minor deviations and achieve high-precision registration.

[0115] Coarse registration is an existing technology, and the specific registration process is as follows:

[0116] The positions of ground control points in the oblique photogrammetry data were accurately identified manually, and the three-dimensional coordinates of these ground control points in the image point cloud coordinate system were measured.

[0117] Identify the location of each ground control point in the ground point cloud and measure the three-dimensional coordinates of these ground control points in the lidar point cloud coordinate system;

[0118] The coarse registration transformation parameters (rotation matrix and translation vector) are calculated based on the three-dimensional coordinates of the ground control points in the image point cloud coordinate system and the three-dimensional coordinates of the ground control points in the lidar point cloud coordinate system.

[0119] Based on the coarse registration transformation parameters, the image point cloud is transformed to obtain a new image point cloud that is roughly aligned with the ground point cloud after coarse registration.

[0120] In registration of complex terrains, the presence of numerous mismatched point pairs can lead to physically unreasonable large-angle rotations when solving for the optimal rotation matrix using traditional ICP algorithms, resulting in iterative divergence or getting trapped in local optima. To address this issue, this embodiment employs an improved ICP algorithm. Its core idea is to set a reasonable range for the Euler angles (or equivalent rotation vectors) corresponding to the rotation matrix calculated in each iteration. This range can be provided by prior knowledge (e.g., the sensor's attitude change between two data acquisitions is not significant) or estimated from the coarse registration results. Then, the Euler angles are constrained based on this reasonable range to avoid unreasonable large-angle rotations.

[0121] In addition to the features of the above embodiments, this embodiment further specifies the steps of fine registration using an improved iterative nearest-point algorithm, specifically including:

[0122] Step 2.31: Use the coarsely registered ground point cloud as the target point cloud and the image point cloud as the source point cloud to be registered.

[0123] Step 2.32: For each point in the source point cloud, find its closest Euclidean distance counterpart in the target point cloud to form the current set of matched point pairs.

[0124] Step 2.33: Construct an objective function based on the currently matched set of point pairs, and solve for the rigid body transformation matrix that minimizes the sum of distances between corresponding points; wherein, the rigid body transformation matrix includes the rotation matrix R. k Translation vector t k The subscript k indicates the current iteration, and the initial value of the rigid body transformation matrix is ​​the transformation matrix obtained from coarse registration. Specifically, the objective function is:

[0125] (1)

[0126] in, Represents the rotation matrix; This represents the i-th point in the source point cloud; Represents the translation vector; This represents the i-th point in the target point cloud; Represents the norm.

[0127] Step 2.34: Based on prior knowledge or coarse registration results, determine the reasonable range of rotation angle variation in the current iteration, specifically including:

[0128] Step 2.341: Initialize the global baseline and range based on the coarse registration results.

[0129] The rotation matrix obtained from coarse registration is converted into Euler angles to obtain a set of initial rotation angle reference values. This rotation angle reference value characterizes the macroscopic rotational deviation to be corrected between the source point cloud and the target point cloud.

[0130] Step 2.342: Take the maximum absolute value of each component of the rotation angle reference value as the reference value: Based on this reference value and a relaxation coefficient β (0.3 ≤ β ≤ 0.7), calculate an initial allowable deviation angle: .

[0131] To ensure rationality, Δθ init Apply boundary constraints to satisfy Δθ min ≤Δθ init ≤Δθ max , where Δθ min and Δθ max These are the preset global minimum allowable deviation angle and global maximum allowable deviation angle (e.g., 0.5° and 10°). Initialize the maximum allowable deviation angle for each axis: Δθ x 0 =Δθ y 0 =Δθ z 0 =Δθ init .

[0132] At the start of the first iteration of fine registration, the mean rotation angle (θ) is... xb 0 ,θ yb 0 ,θ zb 0 Initializing it to 0 means starting fine-tuning from the attitude after coarse registration.

[0133] Step 2.343: Dynamically update the range of change during iteration.

[0134] For the k-th iteration (i.e., the current iteration), its reasonable range of variation (Δθ) x k ,Δθ y k ,Δθ z kThe range is dynamically updated from the previous iteration. Taking the X-axis as an example, Δθ x k =Δθ x k-1 ×γ1, where γ1 represents the first attenuation factor, 0 < γ1 < 1.

[0135] Step 2.344: Determine the mean rotation angle and reasonable range of variation for the current iteration.

[0136] Combining the maximum permissible deviation angle of the current iteration with the average rotation angle constitutes the reasonable range of rotation angle variation in the current iteration:

[0137] θ x k ∈[θ xb k -Δθ x k ,θ xb k +Δθ x k (2)

[0138] θ y k ∈[θ yb k -Δθ y k ,θ yb k +Δθ y k (3)

[0139] θ z k ∈[θ zb k -Δθ z k ,θ zb k +Δθ z k (4)

[0140] Where, θ x k θ y k θ z k θ represents the rotation angle around the X, Y, and Z axes in the current iteration; xb k θ yb k θ zb kThis represents the average rotation angle around the X, Y, and Z axes in the current iteration; its initial value can be obtained from the coarse registration result or set to 0; Δθ x k , Δθ y k , Δθ z k This represents the maximum permissible deviation angle around the X, Y, and Z axes in the current iteration.

[0141] Step 2.35: Convert the rotation matrix obtained in Step 2.33 into Euler angles, and constrain the Euler angles based on the reasonable range of rotation angle variation in the current iteration, specifically as follows:

[0142] θ xconstrained k =max(θ) xb k -Δθ x k min(θ) xb k +Δθ x k ,θ xnew k )) (5)

[0143] θ yconstrained k =max(θ) yb k -Δθ y k min(θ) yb k +Δθ y k ,θ ynew k )) (6)

[0144] θ zconstrained k =max(θ) zb k -Δθ z k min(θ) zb k +Δθ z k ,θ znew k )) (7)

[0145] Where, θ xconstrained k θ yconstrained k θ zconstrained k θ represents the Euler angles constrained by the X, Y, and Z axes in the current iteration.xnew k θ ynew k θ znew k This represents the Euler angles before the constraints on the X, Y, and Z axes in the current iteration, which are obtained by transforming the rotation matrix solved in step 2.33.

[0146] Step 2.36: Convert the constrained Euler angles into a rotation matrix R k,constrained .

[0147] Step 2.37: From the constrained rotation matrix R k,constrained Translation vector t k To construct a new rigid body transformation matrix, the translation vector t k The rigid body transformation matrix obtained from step 2.33 includes the rotation matrix R. k Translation vector t k The specific solution process is based on existing technology.

[0148] Step 2.38: Update the source point cloud according to the new rigid body transformation matrix. The specific update formula is as follows:

[0149] (8)

[0150] in, This represents the i-th point in the updated source point cloud.

[0151] Step 2.39: Determine whether the dynamic convergence condition has been met. If yes, output the final rigid body transformation matrix; otherwise, proceed to step 2.32 for the next iteration.

[0152] In this embodiment, the dynamic convergence condition is:

[0153] RMSE k <τ×α k or |RMSE k -RMSE k-1 |<τ(9)

[0154] Among them, RMSE k RMSE represents the root mean square error (or the change in transformation parameters) of the current iteration. k-1 α represents the root mean square error (or the change in transformation parameters) of the previous iteration; k This represents the dynamic iteration coefficient for the current iteration, whose value decreases as the iteration number k increases, i.e., α. k =α0×γ2 k γ2 represents the second decay factor, 0 < γ2 < 1, α0 represents the initial coefficient, usually set to 1; τ represents the convergence tolerance threshold.

[0155] In this embodiment, rotation angle constraints and dynamic iteration coefficients are innovatively introduced during the registration process, which effectively avoids registration errors caused by complex terrain and a large amount of vegetation, improves registration accuracy and efficiency, and ultimately controls the plane and elevation errors to the centimeter level.

[0156] After obtaining the final rigid body transformation matrix through coarse and fine registration, we can extract pure ground point clouds using laser point cloud filtering. On the other hand, through processes such as triangulation and texture mapping, we can construct a high-precision real-world 3D model that combines the realism of real-world 3D with the ability of laser point clouds to penetrate vegetation, greatly improving the accuracy of subsequent earthwork calculations.

[0157] This embodiment employs a UAV oblique photogrammetry flight scheme and a LiDAR scanning strategy. By leveraging the complementary advantages of these two technologies and an innovative registration algorithm, it successfully solves the industry challenge of inaccurate and incomplete data acquisition in complex terrain and densely vegetated areas. This provides a unique, reliable, and high-precision data foundation for subsequent accurate earthwork calculations, automatic slope design under multi-parameter constraints, and multi-objective optimization algorithms. It is the fundamental guarantee for achieving the intelligent and lean design goals of the entire scheme of this invention.

[0158] In substation layout optimization, the essence of the optimization process is to generate the optimal design scheme by adjusting a series of key decision variables. These decision variables collectively define the three-dimensional morphology of the site, the amount of earthwork, and the cost. Therefore, before constructing the comprehensive objective function, it is necessary to define each decision variable. In this embodiment, the decision variables include the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit, and the earthwork allocation volume.

[0159] Based on the substation site's area and terrain complexity, the earthwork balance calculation area in the 3D model is divided into regular grids. The size of a single grid can be set according to engineering accuracy requirements, such as 10m × 10m or 20m × 20m. Each grid point (i.e., the corner point of the grid) is then numbered, and the design elevation of each grid point l is specified. As an independent decision variable, it is determined by the algorithm within the allowed elevation boundaries during the optimization process. , Adjustments and optimizations will be made within [the specified area]. All The set together represents the optimized site design surface.

[0160] To meet functional requirements such as site drainage, the site is designed as a sloping plane with a single direction, determined by the slope in two directions: the design slope of the site in the east-west direction (or the designated X-axis direction). The design slope of the site in the north-south direction (or the designated Y-axis direction). During the optimization process, the overall slope of the site ( , As a decision variable, its value will directly affect the design elevation of all grids. The overall distribution trend.

[0161] A grid point refers to the corner point of each grid cell. A grid cell is a region enclosed by four adjacent grid points. A slope cell is a computational unit that separately defines the slope surface formed in the boundary region where there is a height difference between the site and the surrounding original ground. A slope consists of multiple slope cells. The location, extent, and shape of a slope cell are entirely determined by the grid points at the site boundary and their design elevations. When the algorithm adjusts the design elevations of these grid points, the slope ratio, height, and shape of the slope also change. Although geometrically dependent, in computation, a slope cell is a special region independent of the internal earthwork grid cells.

[0162] In areas where there is a difference in elevation between the site and its surroundings, slopes will be automatically formed. The boundary area where slopes are required is divided into several slope units, each denoted as (p, q). The slope unit (p, q) represents the block in the p-th row and q-th column. For each slope unit (p, q), its slope height... Defined as the maximum cut / fill height between the designed topography and the natural topography at this unit. Slope height. It is calculated from the design elevation of adjacent grids, but is simultaneously subject to slope safety constraints. Strict restrictions are imposed to ensure the stability of the slope.

[0163] To achieve earthwork balance and optimize transportation costs, refined control of earthwork flow is implemented: grid cells requiring excavation are marked as excavation cell i, and grid cells requiring filling are marked as filling cell j, with earthwork allocation volume... This is defined as the volume of earthwork transported from cut unit i (or the i-th cut unit) to fill unit j (or the j-th fill unit). The optimization algorithm will automatically determine all... The optimal value is obtained, thus forming an earthwork allocation scheme with the lowest total transportation cost.

[0164] The comprehensive objective function of this invention consists of three competing key indicators: earthwork balance objective, economic objective, and safety objective, which are optimized collaboratively through a multi-objective optimization algorithm.

[0165] The earthwork balance objective aims to achieve a balance between cut and fill earthwork within the site, minimizing the borrowing and disposal of earthwork. In addition to the features of the above embodiments, this embodiment further defines the earthwork balance objective as follows:

[0166] (10)

[0167] in, Indicates the earthwork balance target; This represents the excavation volume of the k-th grid cell; This represents the fill volume of the k-th grid cell; N represents the number of grid cells. This represents the penalty coefficient, which is a large positive number. and These represent the total excavation volume and the total fill volume, respectively. , When the excavation and filling volumes are unequal, the earthwork balance target increases, strongly driving the optimization scheme to tend towards earthwork balance.

[0168] The economic objective aims to minimize the total cost of the project, primarily including earthwork excavation and filling costs, earthwork transportation costs, and slope protection costs. In addition to the features of the above embodiments, this embodiment further defines the economic objective as follows:

[0169] (11)

[0170] in, Indicates economic objectives; This represents the cost coefficient per unit of earthwork excavation or filling, expressed in yuan per cubic meter. The unit distance transportation cost coefficient for a unit volume of earthwork, in yuan / (cubic meter·meter). This represents the transportation distance from the i-th cut unit to the j-th fill unit, in meters. This represents the amount of earthwork allocated from the i-th excavation unit to the j-th fill unit, in cubic meters. The slope protection cost coefficient per unit area, in yuan / square meter; The total surface area of ​​the slope is expressed in square meters; m represents the number of cut units; n represents the number of fill units. N = m + n + N0, where N0 represents the number of grid units that require neither cut nor fill.

[0171] The safety objective focuses on slope stability, improving the overall design safety margin by maximizing the safety factor of the most dangerous slope. In addition to the features of the above embodiments, this embodiment further defines the safety objective as follows:

[0172] or (12)

[0173] in, Indicates security objectives; This indicates the safety factor of the slope.

[0174] During the optimization process, the safety factor of all slopes is calculated in real time, with a focus on the slope with the lowest safety factor, thereby improving its safety.

[0175] To ensure the feasibility of the optimized solution in engineering, the present invention also imposes engineering constraints. In addition to the features of the above embodiments, this embodiment further defines the engineering constraints as including geometric constraints, elevation boundary constraints, earthwork balance constraints, and safety constraints.

[0176] The geometric constraint is that the slope between any two adjacent grid points is less than or equal to the maximum allowed slope.

[0177] (13)

[0178] in, , These represent the design elevations of adjacent grid points; Indicates the planar spacing between adjacent grid points; This indicates the maximum permissible slope.

[0179] The elevation boundary constraint is that the design elevation of any grid point lies between the minimum and maximum allowable design elevations, i.e. ,in, , These represent the minimum and maximum allowable design elevations, respectively.

[0180] Earthwork balance constraints are:

[0181] (14)

[0182] in, Indicates the total excavation volume; Indicates the total fill volume; This represents the final looseness coefficient (the volume ratio after compaction, generally taken as 1.03~1.04 for soil). It represents the initial looseness coefficient (the volume expansion ratio after loosening, generally taken as 1.2~1.3 for soil). Indicates the volume of excavated soil; This indicates the amount of earthwork borrowed from outside the company.

[0183] Safety constraints are the slope safety factor Safety factor greater than the specifications or technical requirements Furthermore, the slope height of any slope element must not exceed its maximum allowable height, i.e. .

[0184] In addition to the features of the above embodiments, this embodiment further limits the engineering constraints to include resource constraints, such as limitations on construction resources, equipment, and construction period.

[0185] This invention addresses a complex engineering optimization problem with multiple objectives and constraints. The aim is to find a set of substation layout schemes that achieve the optimal balance among multiple objectives, including earthwork balance, economic cost, and safety. NSGA-II (Non-Dominated Sorting Genetic Algorithm II) exhibits superior performance in handling multi-objective, multi-constraint optimization problems. It outputs a set of Pareto optimal solutions, rather than a single solution, providing decision-makers with multiple preferred options. Therefore, this embodiment further specifies that a non-dominated sorting genetic algorithm is used to solve the comprehensive objective function for multi-objective optimization. Specifically, the NSGA-II algorithm framework, including fast non-dominated sorting, is applied to the substation layout optimization problem. A specific set of decision variables is constructed, including grid design elevation, overall site slope, slope height, and earthwork allocation, and a comprehensive objective function integrating economic efficiency, earthwork balance, and safety, along with a series of engineering constraints, is integrated.

[0186] like Figure 3 As shown, a non-dominated sorting genetic algorithm is used to solve the comprehensive objective function for multi-objective optimization, including:

[0187] Step 5.1: Encode the decision variable set into chromosomes using real numbers and generate the initial population.

[0188] This embodiment uses real number encoding, concatenating all decision variables sequentially to form a one-dimensional vector, which is the chromosome, representing a complete solution. Therefore, an individual chromosome can be represented as:

[0189] (15)

[0190] Here, M represents the number of grid points. A certain number of initial individuals (e.g., 100) are randomly generated to form the initial population. To ensure the quality and feasibility of the initial solution, a heuristic method (e.g., a small perturbation based on the average elevation of the original terrain) is used instead of completely random generation to improve convergence efficiency.

[0191] Step 5.2: Calculate the function values ​​of each individual in the population corresponding to the earthwork balance target, economic target, and safety target, and verify their satisfaction with the engineering constraints.

[0192] For each individual (scheme) in the population, it is necessary to calculate its performance on the earthwork balance objective, economic objective, and safety objective, and verify whether it meets all constraints.

[0193] Earthwork balance objective (minimization) ): Calculate the absolute difference between cut and fill for each individual The ideal value is 0.

[0194] Economic objectives (minimization) ): Calculate the total earthwork volume for each individual. (equal to the volume of excavation) +fill volume ) and earthwork transportation costs (transport volume) ×Transportation distance).

[0195] Security goals (maximization) The safety factor of the most dangerous slope can be calculated (using methods such as the simplified Bishop method), or the maximum slope height can be used as a proxy. The simplified Bishop method is the current technology for calculating the slope safety factor.

[0196] Check whether each individual violates all constraints. These constraints are usually handled by the penalty function method (converting the degree of constraint violation into a numerical penalty, which is added to the objective function to make it difficult for infeasible solutions to survive).

[0197] Step 5.3: Perform a fast non-dominated ordination on each individual in the population to determine its non-dominated rank, and calculate the crowding distance for individuals within the same non-dominated rank.

[0198] All individuals in the population are ranked using fast non-dominated sorting.

[0199] Rank 1 consists of non-dominated solutions that cannot be surpassed by any other individual (i.e., Pareto optimal solutions). These non-dominated solutions are better than all other solutions in at least one objective and are not worse in other objectives.

[0200] Second Rank: Solutions dominated by solutions in the first rank, but not by solutions in other ranks.

[0201] This process continues until all individuals are ranked to determine their non-dominated rank. A smaller Rank value indicates a better solution. The Rank value refers to the level of a non-dominated solution; a Rank 1 solution is a non-dominated solution in the current population, while a Rank 2 solution is dominated by Rank 1 solutions but not by other solutions.

[0202] Within the same non-dominated layer (same rank), to maintain population diversity, the density of each individual around it needs to be calculated, i.e., the crowding distance. The crowding distance reflects the average distance between a solution and its neighboring solutions in the target space. The larger the crowding distance, the more open the area around the solution, indicating good diversity, and it should be preserved.

[0203] Step 5.4: Generate offspring populations based on non-dominant level and crowding distance through selection, crossover, and mutation operations.

[0204] Two individuals are randomly selected from the population using a binary tournament selection method, prioritizing individuals with higher non-dominant rank (lower rank value); if two individuals have the same rank, the individual with a larger crowding distance is selected (to preserve diversity); the selected individual enters the mating pool and becomes the parent generation.

[0205] Two parent individuals are randomly selected from the mating pool, and a portion of their genes (decision variables) are exchanged with a certain probability (e.g., 0.9) to produce new offspring. This embodiment uses simulated binary crossover (SBX), which is particularly suitable for real-number encoding, can produce offspring similar to the parents, and has good convergence.

[0206] With a small probability (e.g., 0.1), one or more gene values ​​in offspring individuals are randomly changed (e.g., the design elevation of a grid point) to introduce new genes, achieve mutation, and escape local optima. In this embodiment, polynomial mutation is used.

[0207] Step 5.5: Merge the parent and offspring populations, and select a new generation of population from the merged populations based on non-dominated sorting and crowding distance calculation.

[0208] The parent population (size Q) and the newly generated offspring population (size Q) are merged to form a merged population of size 2Q. This merged population of size 2Q is then re-sorted for non-dominated ranking and crowding distance calculation. Individuals with the smallest Rank value are selected first, and this process continues until Q individuals are selected, forming a new generation population.

[0209] The elite retention strategy ensures that the best individuals (elites) from the previous generation are not discarded, thereby accelerating convergence and ensuring that the algorithm converges to the Pareto optimal front with probability 1.

[0210] Step 5.6: Determine if the termination condition is met. If yes, output the Pareto optimal solution set consisting of the individuals with the highest non-dominant rank in the final population. If no, proceed to step 5.2.

[0211] In this embodiment, the termination condition is that the number of iterations reaches the set maximum number of iterations (e.g., 500), or the solution set satisfies the convergence condition. Specifically, convergence is considered achieved when the rate of change of the generational distance of the Pareto optimal solution set over several consecutive generations is less than a preset tolerance threshold (e.g., 0.001). Generational distance (GD) is an indicator that measures the difference between the currently found solution set and the true global optimal solution set. Generational distance is a technical term, and its calculation is based on existing technology.

[0212] After the algorithm terminates, all individuals with Rank 1 in the final population (non-dominated solutions) constitute the Pareto optimal frontier. These solutions represent the set of best compromise solutions among multiple objectives such as earthwork balance, economy, and safety, which cannot be further optimized.

[0213] This invention does not simply output a solution, but provides a set of Pareto optimal solutions. Decision-makers can select the final optimal implementation scheme from the Pareto optimal solution set based on actual engineering preferences (e.g., more focus on cost or more focus on safety) using decision-making tools such as TOPSIS (Topology of Ideal Solutions) or AHP (Analytic Hierarchy Process).

[0214] Example 2

[0215] like Figure 4 As shown, the earthwork balance calculation system disclosed in this embodiment includes an acquisition unit, a fusion processing unit, a grid generation unit, a definition and construction unit, and a solution and selection unit.

[0216] The acquisition unit is used to acquire oblique photography data and lidar point cloud data of the target area;

[0217] The fusion processing unit is used to fuse oblique photogrammetry data and lidar point cloud data to generate a high-precision real-scene 3D model.

[0218] Mesh generation units are used to divide the real-world 3D model into meshes, resulting in a meshed model for optimization calculations;

[0219] Define and construct units to define a set of decision variables based on a gridded model, including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit, and the earthwork allocation volume. Construct a comprehensive objective function that includes earthwork balance objectives, economic objectives, and safety objectives based on the set of decision variables.

[0220] The solution and selection unit is used to perform multi-objective optimization on the comprehensive objective function under preset engineering constraints, obtain the Pareto optimal solution set, and select the final implementation scheme from the Pareto optimal solution set.

[0221] In some specific embodiments of the present invention, the earthwork balance calculation system can be combined with the features of the earthwork balance calculation method in Embodiment 1 of the present invention, and vice versa, which will not be repeated here.

[0222] Example 3

[0223] This invention also provides an electronic device, such as... Figure 5 As shown, the electronic device includes: a memory, a processor, and a computer program / instructions stored in the memory. The processor executes the computer program / instructions to implement the steps in the earthwork balance calculation method of Embodiment 1 of the present invention.

[0224] Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) or loaded from a storage portion into random access memory (RAM). The processor can be a multi-core processor or may contain multiple processors. In some embodiments, the processor may include a general-purpose main processor and one or more specialized coprocessors, such as a central processing unit, graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for device operation are also stored in RAM. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0225] The processor and memory described above are used together to execute programs / instructions stored in the memory. When the program / instructions are executed by the computer, they can implement the methods, steps, or functions described in the above embodiments.

[0226] Although not shown, embodiments of the present invention also provide a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps in the earthwork balance calculation method of embodiment 1 of the present invention.

[0227] Readable storage media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0228] Although not shown, embodiments of the present invention also provide a computer program product, including: a computer program / instructions that, when executed by a processor, implement the steps in the earthwork balance calculation method of embodiment 1 of the present invention.

[0229] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0230] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for calculating earthwork balance, characterized in that, The method includes: Acquire oblique photography data and lidar point cloud data of the target area; The oblique photography data and lidar point cloud data are fused to generate a high-precision real-scene 3D model. The real-world 3D model is divided into meshes to obtain a meshed model for optimization calculations; Based on the gridded model, a set of decision variables is defined, including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit, and the earthwork allocation. A comprehensive objective function containing earthwork balance, economic and safety objectives is constructed based on the set of decision variables. Under the preset engineering constraints, the comprehensive objective function is solved by multi-objective optimization to obtain the Pareto optimal solution set, and the final implementation scheme is selected from the Pareto optimal solution set; The earthwork balance target is as follows: ； in, Indicates the earthwork balance target; This represents the excavation volume of the k-th grid cell; This represents the fill volume of the k-th grid cell; N represents the number of grid cells. Indicates the penalty coefficient; and These represent the total excavation volume and the total fill volume, respectively. , ; The economic objective is as follows: ； in, Indicates economic objectives; This represents the cost coefficient per unit of earthwork excavation or filling. This represents the unit distance transportation cost coefficient for a unit volume of earthwork. This represents the transportation distance from the i-th excavation unit to the j-th fill unit; This represents the amount of earthwork allocated from the i-th excavation unit to the j-th fill unit; This represents the slope protection cost coefficient per unit area. The total surface area of ​​the slope is represented by m; the number of cut units is represented by n; and the number of fill units is represented by n. The security objective is: or ; in, Indicates security objectives; This indicates the safety factor of the slope.

2. The earthwork balance calculation method according to claim 1, characterized in that, The acquisition of oblique photography data and lidar point cloud data of the target area includes: The target area is surveyed, and based on the terrain complexity and vegetation coverage, a drone oblique photography flight plan and a lidar scanning strategy are planned in a coordinated manner. The UAV oblique photography flight scheme includes a grid-shaped flight path or a five-lens camera acquisition scheme, ensuring that the overlap between the flight path and the lateral direction is not less than 60% to obtain ground texture information and macroscopic geometric shape; the lidar scanning strategy includes selecting at least one of airborne, ground-based, or handheld SLAM methods for scanning according to data gap filling requirements, and setting the scanning distance, scanning density, and scanning angle to actively penetrate vegetation gaps and obtain high-precision three-dimensional point clouds of the ground surface and hidden areas. Based on the UAV oblique photography flight scheme, the oblique photography data is acquired; based on the lidar scanning strategy, the lidar point cloud data is acquired.

3. The earthwork balance calculation method according to claim 1, characterized in that, The oblique photography data and lidar point cloud data are fused, including: Aerial triangulation and dense matching are performed on the oblique photogrammetry data to generate densely matched image point clouds; The lidar point cloud data is denoised, filtered, and classified to extract clean ground point clouds; The image point cloud and the ground point cloud are coarsely and finely registered to generate fused point cloud data. Based on the fused point cloud data, a high-precision real-scene 3D model is constructed.

4. The earthwork balance calculation method according to claim 3, characterized in that, A refined iterative nearest-point algorithm is used for fine registration, including: The ground point cloud after coarse registration is used as the target point cloud, and the image point cloud is used as the source point cloud to be registered. For each point in the source point cloud, find its closest Euclidean distance counterpart in the target point cloud to form the current set of matched point pairs; Based on the currently matched set of point pairs, an objective function is constructed, and the rigid body transformation matrix that minimizes the sum of the distances between corresponding points is solved; wherein, the rigid body transformation matrix includes a rotation matrix and a translation vector, and the initial value of the rigid body transformation matrix is ​​the transformation matrix obtained by coarse registration; Based on prior knowledge or coarse registration results, determine the reasonable range of rotation angle variation in the current iteration; The rotation matrix is ​​converted into Euler angles, and the Euler angles are constrained based on the reasonable range of variation of the rotation angles in the current iteration. Convert the constrained Euler angles into rotation matrices; A new rigid body transformation matrix is ​​formed by the constrained rotation matrix and the translation vector; Update the source point cloud based on the new rigid body transformation matrix; Determine whether the dynamic convergence condition has been met. If yes, output the final rigid body transformation matrix; otherwise, proceed to the step of forming the current set of matched point pairs and perform the next iteration.

5. The earthwork balance calculation method according to claim 1, characterized in that, The engineering constraints include geometric constraints, elevation boundary constraints, earthwork balance constraints, and safety constraints; The geometric constraint is that the slope between any two adjacent grid points is less than or equal to the maximum allowable slope; the elevation boundary constraint is that the design elevation of any grid point is between the minimum and maximum allowable design elevations; the earthwork balance constraint is: ； in, Indicates the total excavation volume; Indicates the total fill volume; Indicates the final looseness coefficient; Indicates the initial looseness coefficient; Indicates the volume of excavated soil; Indicates the amount of earthwork borrowed; The safety constraint is that the slope safety factor is greater than the safety factor required by the specifications or technical requirements, and the slope height of any slope unit shall not exceed its maximum allowable height.

6. The earthwork balance calculation method according to any one of claims 1 to 5, characterized in that, A non-dominated sorting genetic algorithm is used to solve the comprehensive objective function for multi-objective optimization, including: The set of decision variables is encoded into chromosomes using real numbers, and an initial population is generated. Calculate the function value of each individual in the population corresponding to the earthwork balance target, economic target and safety target, and verify its satisfaction with the engineering constraints; Perform a fast nondominated ordination on each individual in the population to determine its nondominated rank, and calculate the crowding distance for individuals within the same nondominated rank. Based on the non-dominant rank and crowding distance, offspring populations are generated through selection, crossover, and mutation operations; The parent and offspring populations are merged, and a new generation of population is selected from the merged populations based on non-dominated ranking and crowding distance calculations. Determine if the termination condition is met. If yes, output the Pareto optimal solution set composed of the individuals with the highest non-dominant level in the final population. If no, proceed to the step of calculating the function value of each individual in the population corresponding to the earthwork balance target, economic target, and safety target, and perform the next iteration loop.

7. An earthwork balance calculation system, characterized in that, The system includes: The acquisition unit is used to acquire oblique photography data and lidar point cloud data of the target area; The fusion processing unit is used to fuse the oblique photography data and the lidar point cloud data to generate a high-precision real-scene 3D model. Mesh division unit, used to divide the real scene 3D model into meshes to obtain a meshed model for optimization calculation; Define and construct a unit, which is used to define a set of decision variables based on the gridded model, including the design elevation of each grid point, the overall slope of the site, the slope height of each slope unit and the earthwork allocation, and construct a comprehensive objective function including earthwork balance objective, economic objective and safety objective based on the set of decision variables; The solution and selection unit is used to perform multi-objective optimization on the comprehensive objective function under preset engineering constraints, obtain the Pareto optimal solution set, and select the final implementation scheme from the Pareto optimal solution set; The earthwork balance target is as follows: ； in, Indicates the earthwork balance target; This represents the excavation volume of the k-th grid cell; This represents the fill volume of the k-th grid cell; N represents the number of grid cells. Indicates the penalty coefficient; and These represent the total excavation volume and the total fill volume, respectively. , ; The economic objective is as follows: ； in, Indicates economic objectives; This represents the cost coefficient per unit of earthwork excavation or filling. This represents the unit distance transportation cost coefficient for a unit volume of earthwork. This represents the transportation distance from the i-th excavation unit to the j-th fill unit; This represents the amount of earthwork allocated from the i-th excavation unit to the j-th fill unit; This represents the slope protection cost coefficient per unit area. The total surface area of ​​the slope is represented by m; the number of cut units is represented by n; and the number of fill units is represented by n. The security objective is: or ; in, Indicates security objectives; This indicates the safety factor of the slope.

8. An electronic device comprising a memory, a processor, and a computer program / instructions stored in the memory, characterized in that, The processor executes the computer program / instructions to implement the steps in the earthwork balance calculation method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps in the earthwork balance calculation method as described in any one of claims 1 to 6.

Images