Pit volume intelligent detection system based on three-dimensional laser scanning technology

Abstract

The invention discloses an innovative pit body volume detection system based on a three-dimensional laser scanning technology. Acquiring the volume point cloud information of the pit body by adopting a three-dimensional laser scanning technology; the volume is calculated through double-engine cross validation, two algorithms are used for cross validation, and it is ensured that the volume calculation result relative error meets the requirement; comprising the steps that a Matlab layered integral algorithm is adopted, specifically, discrete layered volume integration is achieved through elevation layered cutting, boundary polygon generation of a polyshape function and AlphaShape geometric reconstruction; and a Geomagic curved surface reconstruction algorithm: generating a watertight grid model based on an NURBS curved surface technology, and completing continuous volume calculation. According to the method, a non-contact measurement system is established, so that the measurement efficiency is greatly improved; a three-dimensional point cloud digital processing platform is created, full-process digitalization of scanning data automatic registration, boundary recognition and volume calculation and real-time visual output of measurement results are achieved, secondary errors caused by manual intervention are reduced, the project acceptance data traceability requirement is met, and the measurement comprehensive error rate is greatly reduced.

Description

Technical Field

[0001] The present invention belongs to the field of surveying and mapping technology, in particular to the field of laser digital engineering surveying and mapping technology, and specifically relates to an intelligent pit volume detection system based on three-dimensional laser scanning technology. Background Art

[0002] In the fields of civil engineering, environmental governance, and mine restoration, the accurate measurement of pit volume is a key link in engineering quality control. Traditional measurement methods generally use the water volume replacement method, and its specific implementation process is: laying a geomembrane in the pit to be measured to form a closed container, then injecting water, and inferring the pit volume by measuring the amount of injected water. This method has significant technical defects in actual application: first, geomembranes are prone to puncture and damage under complex geological conditions, resulting in liquid leakage and serious distortion of measurement data; second, the measurement process involves multiple steps such as material transportation, membrane laying, water injection and drainage, and the operation process is cumbersome and time-consuming for several hours; third, affected by external factors such as ambient temperature and evaporation, the water volume is prone to measurement deviations, especially in open-air operation scenarios where the error rate is relatively large.

[0003] Therefore, this field urgently needs an intelligent pit volume detection technology solution that can effectively improve the accuracy and efficiency of pit volume measurement, realize the intelligence of the measurement process, and meet the needs of high-precision engineering measurement. Summary of the Invention

[0004] The purpose of the present invention is to overcome the technical defects of the existing water volume replacement method and provide an innovative pit volume detection system based on three-dimensional laser scanning technology.

[0005] The present invention is achieved through the following technical solutions:

[0006] An intelligent pit volume detection system based on 3D laser scanning technology is characterized by: using 3D laser scanning technology to obtain pit volume point cloud information; using dual-engine cross-validation to calculate volume, and using two algorithms for cross-validation to ensure that the relative error of the volume calculation results meets the requirements; including:

[0007] Matlab hierarchical integration algorithm: Discrete hierarchical volume integration is achieved through elevation hierarchical cutting, boundary polygon generation polyshape function and AlphaShape geometric reconstruction;

[0008] Geomagic surface reconstruction algorithm: Generates a watertight mesh model based on NURBS surface technology and completes continuous volume calculation.

[0009] The present invention provides an intelligent pit volume detection system based on three-dimensional laser scanning technology, which specifically includes the following methods: (1) deployment of a three-dimensional laser scanning system and acquisition of volume point cloud data; (2) real-time fusion processing of point cloud data; (3) preprocessing of point cloud data using RiscanPro software; (4) layered volume calculation based on Matlab; and (5) modeling volume calculation based on Geomagic.

[0010] Further, the deployment of the three-dimensional laser scanning system and the volume point cloud data acquisition in the method (1) include: using a multi-line array pulsed three-dimensional laser scanner, integrating a high-precision positioning module, and adapting to panoramic scanning of complex terrain and high-steep structures; coordinated control of the horizontal scanning rate and angular resolution to achieve high-density point cloud data acquisition to meet the subsequent modeling accuracy requirements.

[0011] Further, the method (2) real-time fusion processing of point cloud data includes: based on the collected point cloud data, using the registration technology to achieve spatiotemporal alignment; performing point cloud downsampling operations to retain key morphological features while reducing data redundancy; using a spatial filtering algorithm combined with index structure optimization to efficiently eliminate outlier noise points and generate a standardized three-dimensional coordinate matrix; the processing flow follows the technical logic of point cloud splicing, registration, and noise reduction, and the output data is directly used for high-precision modeling and volume calculation.

[0012] Further, the method (3) pre-processes the point cloud data using Riscan Pro software, including:

[0013] (31) The point cloud data collected at multiple stations are globally registered using the Multi Station Adjustment function combined with the ICP algorithm and unified into the same coordinate system;

[0014] (32) Perform voxel grid downsampling to reduce data redundancy while preserving the pit morphological characteristics, and use spatial mean filtering and reflectivity threshold screening to remove vegetation and environmental noise in real time;

[0015] (33) The point cloud topology structure is reconstructed through the built-in spatial index optimization tool of the software to generate a standardized three-dimensional coordinate matrix. The output format supports .las or .txt, providing a high-precision data foundation for subsequent modeling.

[0016] Further, the method (4) comprises the following steps based on Matlab layered volume calculation:

[0017] (41) Loading and thinning point cloud data; Select a point cloud file (.txt format) through the interactive interface and convert it into a Matlab point cloud object using the pointCloud function; Use the voxel grid thinning method (pcdownsample function) and set the voxel size to 0.005 meters to reduce data redundancy while retaining morphological features;

[0018] (42) Layered cutting and local geometry reconstruction; the point cloud is divided into 50 layers along the elevation direction, and the height of each layer is dynamically calculated according to the depth of the pit; the point cloud of the top 5% of the height of each layer is extracted, and the boundary polygon polyshape function is generated by polar angle sorting; regular grid points are generated within the boundary with a step size of 0.05 meters; internal points are screened and interpolated to generate a closed top surface; the original layer point cloud, the interpolated top surface and the bottom surface of the previous layer are integrated to construct the complete layer geometry;

[0019] (43) AlphaShape layered integration and volume calculation; call the alphaShape function for each layer of data, and optimize the surface fitting effect by adjusting the α parameter; calculate the volume layer by layer and accumulate it, and finally output the total volume of the pit.

[0020] Further, the method (5) includes the following steps based on the Geomagic modeling volume calculation:

[0021] (51) Import pre-processed point cloud data and maintain 100% sampling rate for full data modeling;

[0022] (52) Execute the normal automatic repair operation, and perform the view orientation normal → automatic orientation → selective flip normal in sequence;

[0023] (53) Surface encapsulation is performed using a point spacing parameter of 0.01 m, and the “delete widget” function is simultaneously activated to remove invalid data;

[0024] (54) Apply the "Fill Single Hole" algorithm to repair model defects and improve geometric continuity through the "Fill All" operation;

[0025] (55) Call the volume analysis module to calculate the pit volume and finally output the pit volume.

[0026] The present invention achieves the following technical objectives and benefits:

[0027] The method of the present invention establishes a non-contact measurement system and eliminates liquid leakage errors caused by geomembrane damage through three-dimensional laser scanning technology. In some cases, it can achieve millimeter-level precision reconstruction of the pit shape; at the same time, it greatly shortens the time required for the traditional measurement process and greatly improves measurement efficiency.

[0028] The method of the present invention creates a three-dimensional point cloud digital processing platform, realizes the full process digitization of automatic alignment of scanning data, boundary recognition and volume calculation, and outputs the measurement results in real-time visualization, reducing secondary errors caused by human intervention.

[0029] The method of the present invention greatly improves the measurement accuracy and operation efficiency; three-dimensional laser scanning technology is used to obtain pit volume data, and at the same time, the alphaShape algorithm and NURBS surface reconstruction dual computing engines are cross-validated, and the final volume calculation result has a relatively small error, and in some cases can meet the millimeter-level accuracy standard; three-dimensional laser scanning technology is used to replace the original water volume replacement method, which greatly shortens the time required for data acquisition, and creates a three-dimensional point cloud digital processing platform to achieve accurate alignment of scanned data, greatly shortening the time required for traditional processes and greatly improving measurement efficiency.

[0030] The method of the present invention enhances the visualization and traceability of the point cloud data processing process; the present invention provides Matlab code-level calculation logs and Geomagic model reconstruction records, and generates a three-dimensional digital model to meet the data traceability requirements of engineering acceptance.

[0031] The method of the present invention realizes fully automatic and high-precision measurement of pit volume through the scanning-processing dual-stage technical innovation, greatly shortens the single operation cycle, greatly reduces the comprehensive measurement error rate, and has higher superiority than traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a site map of point cloud data obtained using three-dimensional laser scanning technology according to an embodiment of the present invention;

[0033] Figure 2 This is the point cloud model after riscan pro preprocessing is completed in an embodiment of the present invention;

[0034] Figure 3 It is a MATLAB point cloud sparse graph according to an embodiment of the present invention;

[0035] Figure 4 This is a Geomagic point cloud coloring model diagram according to an embodiment of the present invention;

[0036] Figure 5 This is a Geomagic colored polygon mesh model diagram according to an embodiment of the present invention;

[0037] Figure 6 This is a three-dimensional solid model diagram preprocessed according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described below in conjunction with specific embodiments. The specific embodiments are further explanations of the principles of the present invention and do not limit the present invention in any way. Technologies that are the same or similar to the present invention do not exceed the scope of protection of the present invention.

[0039] The present invention provides an intelligent pit volume detection system based on three-dimensional laser scanning technology, which includes the following methods: three-dimensional laser scanning system deployment and point cloud data acquisition; real-time fusion processing of point cloud data; pre-processing of point cloud data using Riscan Pro software; layered volume calculation based on Matlab; and modeled volume calculation based on Geomagic.

[0040] (1) 3D laser scanning system deployment and point cloud data acquisition.

[0041] A multi-line pulsed 3D laser scanner, integrated with a high-precision positioning module, adapts to the panoramic scanning needs of complex terrain and steep structures. The coordinated control of horizontal scanning rate and angular resolution enables high-density point cloud acquisition, meeting the accuracy requirements of subsequent modeling.

[0042] (2) Real-time fusion processing of point cloud data.

[0043] Based on the collected point cloud data, registration technology is used to achieve spatiotemporal alignment. Point cloud downsampling is performed to retain key morphological features while reducing data redundancy. A spatial filtering algorithm combined with index structure optimization is used to efficiently eliminate outlier noise points and generate a standardized 3D coordinate matrix. The processing flow strictly adheres to the technical logic of point cloud stitching, registration, and noise reduction, ensuring that the output data can be directly used for high-precision modeling and volume calculation.

[0044] (3) Use Riscan Pro software to preprocess the point cloud data.

[0045] First, the Multi Station Adjustment function combined with the ICP algorithm is used to globally align the point cloud data collected at multiple sites and unify them into the same coordinate system. Voxel grid downsampling is then performed to reduce data redundancy while retaining the pit morphological characteristics. Spatial mean filtering and reflectivity threshold screening are used to remove vegetation and environmental noise in real time. Finally, the software's built-in spatial index optimization tool is used to reconstruct the point cloud topology and generate a standardized 3D coordinate matrix. The output format supports .las or .txt, providing a high-precision data foundation for subsequent modeling.

[0046] (4) Layered volume calculation based on Matlab, specifically including:

[0047] ① Loading and thinning point cloud data: Select a point cloud file (.txt format) through the interactive interface and convert it into a Matlab point cloud object using the pointCloud function. Use the voxel grid thinning method (pcdownsample function) and set the voxel size to 0.005 meters to reduce data redundancy while preserving morphological features.

[0048] ② Layered cutting and local geometric reconstruction: Divide the point cloud into 50 equal layers along the elevation direction, and the height of each layer is dynamically calculated according to the pit depth; extract the point cloud of the top 5% of the height of each layer (generating a boundary polygon by polar angle sorting (polyshape function); generate a regular grid of points within the boundary (step size 0.05 meters), filter the internal points and interpolate to generate a closed top surface; integrate the original layer point cloud, the interpolated top surface and the bottom surface of the previous layer to construct the complete layer geometry.

[0049] ③AlphaShape layered integration and volume calculation; call the alphaShape function for each layer of data, and optimize the surface fitting effect by adjusting the α parameter; calculate the volume layer by layer and accumulate it, and finally output the total volume of the pit.

[0050] (5) Model-based volume calculation based on Geomagic, including:

[0051] ① Import pre-processed point cloud data and maintain 100% sampling rate for full data modeling;

[0052] ② Perform the normal automatic repair operation, and follow the steps of view orientation normal → automatic orientation → selective flip normal;

[0053] ③Use the parameter of 0.01m point spacing for surface packaging, and simultaneously activate the "delete small components" function to eliminate invalid data;

[0054] ④ Use the "Fill Single Hole" algorithm to repair model defects and use the "Fill All" operation to improve geometric continuity;

[0055] ⑤ Call the volume analysis module to calculate the pit volume and finally output the pit volume.

[0056] The following describes in detail the implementation process of the intelligent pit volume detection system based on three-dimensional laser scanning technology of the present invention.

[0057] (1) 3D laser scanning system deployment:

[0058] Three ground scanning stations are deployed around the pit, and a multi-line array pulsed 3D laser scanner is used for data acquisition. For areas with complex terrain (such as steep slopes and vegetation-covered areas), a collaborative scanning strategy of three stations is used to cover the pit surface and base from different directions, solving the problem of data loss caused by field of view obstruction at a single station. Figure 1 As shown, Figure 1 The image shows the scanning coverage of the pit surface and base from one of the three stations. A coordinated scanning strategy across the three stations covers the pit surface and base from different angles, resolving data loss issues caused by obstructed field of view at a single station. Station spacing is dynamically adjusted based on the actual pit size, and global coordinate registration technology is used to eliminate blind spots. Reflectivity threshold screening removes vegetation noise and environmental interference in real time, ensuring the validity of the surface point cloud.

[0059] (2) Point cloud data collection and real-time fusion processing:

[0060] Start multi-site synchronous scanning to obtain point cloud data of the pit surface and base. After scanning is completed, perform the following processing:

[0061] a. Multi-station registration: Use the MultiStationAdjustment function in Riscan Pro software, combined with the ICP (Iterative Closest Point) algorithm, to unify the data of each station into the global coordinate system. Figure 2 As shown in the figure, this figure shows the distribution and fusion of data in a unified coordinate system after multi-site registration.

[0062] b. Data denoising and optimization: Use voxel grid downsampling (voxel size 5mm) to reduce data redundancy, combine spatial mean filtering with kd-tree index structure to eliminate outlier noise points, and generate a standardized three-dimensional coordinate matrix. Figure 2 As shown in the figure, the point cloud data after noise reduction and optimization is clearer, the interference data with high reflectivity has been effectively eliminated, and the point cloud model finally generated is more accurate.

[0063] (3) Layered volume calculation (Matlab platform):

[0064] a. Point cloud loading and preprocessing: import the point cloud file (.txt format) through the interactive interface, convert it into a Matlab point cloud object, and perform voxel grid thinning (step size 0.005m). Figure 3 As shown, Figure 3 The point cloud data after voxel grid thinning is displayed, which reduces data redundancy and improves processing efficiency.

[0065] b. Layered cutting and geometric reconstruction: The point cloud is divided into 50 equal layers along the elevation direction, and the height of each layer is dynamically calculated. The top 5% of the point cloud is extracted to generate a boundary polygon (polyshape function), and the regular grid top surface is generated by interpolation within the boundary.

[0066] c. AlphaShape integral calculation: Call the alphaShape function (α = 0.06) for each layer of data, integrate the original layer point cloud, the interpolated top surface and the bottom surface of the previous layer, calculate the volume layer by layer and accumulate them, and output the total volume of the pit.

[0067] (4) Modeling volume calculation (Geomagic platform):

[0068] a. Data import and packaging: Import the pre-processed point cloud into Geomagic Studio, maintain 100% sampling rate for full data modeling, and use the point spacing parameter of 0.01m to generate the triangular mesh surface; as shown in the attached Figure 4 As shown, Figure 4 The point cloud shading model after importing into Geomagic Wrap is shown to help visualize the distribution and details of the point cloud data.

[0069] b. Model repair and optimization: Perform automatic normal repair, eliminate holes through "Fill single hole" and "Fill all" operations to ensure the geometric continuity of the model; as shown in the attached Figure 5 As shown, Figure 5 The repaired polygon mesh model is shown, demonstrating the geometric continuity and integrity of the model after repair.

[0070] c. Volume analysis and verification: Call the volume analysis module, calculate the pit volume based on NURBS surface reconstruction technology, and cross-verify with the Matlab calculation results. Figure 6 As shown, Figure 6 The final pre-processed 3D solid model was displayed to verify the accuracy and consistency of the model.

Claims

1. An intelligent pit volume detection system based on 3D laser scanning technology, characterized by: Use 3D laser scanning technology to obtain pit volume point cloud information; use dual-engine cross-validation to calculate volume, and cross-validate two algorithms to ensure that the relative error of the volume calculation results meets the requirements; including: Matlab hierarchical integration algorithm: Discrete hierarchical volume integration is achieved through elevation hierarchical cutting, boundary polygon generation polyshape function and AlphaShape geometric reconstruction; Geomagic surface reconstruction algorithm: Generates a watertight mesh model based on NURBS surface technology and completes continuous volume calculation.

2. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 1 is characterized in that Specifically, the following methods are included: (1) deployment of a 3D laser scanning system and volumetric point cloud data acquisition; (2) real-time fusion processing of point cloud data; (3) preprocessing of point cloud data using Riscan Pro software; (4) layered volume calculation based on Matlab; and (5) model-based volume calculation based on Geomagic.

3. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 2 is characterized by: The deployment of the three-dimensional laser scanning system and the volume point cloud data acquisition in the method (1) include: using a multi-line array pulsed three-dimensional laser scanner, integrating a high-precision positioning module, and adapting to panoramic scanning of complex terrain and high-steep structures; coordinated control of the horizontal scanning rate and angular resolution to achieve high-density point cloud data acquisition to meet the subsequent modeling accuracy requirements.

4. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 2 is characterized by: The method (2) includes the following steps: based on the collected point cloud data, using registration technology to achieve spatiotemporal alignment; performing point cloud downsampling operations to retain key morphological features while reducing data redundancy; using a spatial filtering algorithm combined with index structure optimization to efficiently eliminate outlier noise points and generate a standardized three-dimensional coordinate matrix; the processing flow follows the technical logic of point cloud splicing, registration, and noise reduction, and the output data is directly used for high-precision modeling and volume calculation.

5. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 2 is characterized by: The method (3) uses Riscan Pro software to pre-process the point cloud data, including: (31) The point cloud data collected at multiple stations are globally registered using the Multi Station Adjustment function combined with the ICP algorithm and unified into the same coordinate system; (32) Perform voxel grid downsampling to reduce data redundancy while preserving the pit morphological characteristics, and use spatial mean filtering and reflectivity threshold screening to remove vegetation and environmental noise in real time; (33) The point cloud topology structure is reconstructed through the built-in spatial index optimization tool of the software to generate a standardized three-dimensional coordinate matrix. The output format supports .las or .txt, providing a high-precision data foundation for subsequent modeling.

6. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 2 is characterized by: The method (4) comprises the following steps based on Matlab-based layered volume calculation: (41) Loading and thinning point cloud data; Select a point cloud file (.txt format) through the interactive interface and convert it into a Matlab point cloud object using the pointCloud function; Use the voxel grid thinning method (pcdownsample function) and set the voxel size to 0.005 meters to reduce data redundancy while retaining morphological features; (42) Layered cutting and local geometry reconstruction; the point cloud is divided into 50 layers along the elevation direction, and the height of each layer is dynamically calculated according to the depth of the pit; the point cloud of the top 5% of the height of each layer is extracted, and the boundary polygon polyshape function is generated by polar angle sorting; regular grid points are generated within the boundary with a step size of 0.05 meters; internal points are screened and interpolated to generate a closed top surface; the original layer point cloud, the interpolated top surface and the bottom surface of the previous layer are integrated to construct the complete layer geometry; (43) AlphaShape layered integration and volume calculation; call the alphaShape function for each layer of data, and optimize the surface fitting effect by adjusting the α parameter; calculate the volume layer by layer and accumulate it, and finally output the total volume of the pit.

7. The intelligent pit volume detection system based on 3D laser scanning technology according to claim 2 is characterized by: The method (5) comprises the following steps based on the volume calculation of Geomagic modeling: (51) Import pre-processed point cloud data and maintain 100% sampling rate for full data modeling; (52) Performing normal automatic repair operation, in order of view orientation normal → automatic orientation → selective flip normal; (53) Surface encapsulation is performed using a point spacing parameter of 0.01 m, and the "delete widget" function is activated simultaneously to eliminate invalid data; (54) Apply the "Fill Single Hole" algorithm to repair model defects and improve geometric continuity through the "Fill All" operation; (55) Call the volume analysis module to calculate the pit volume and finally output the pit volume.

Images