LiDAR-based inclined keel bulkhead construction precision management and control method and system

Abstract

The embodiment of the application provides a LiDAR-based inclined keel partition wall construction precision management and control method and system, the method comprises the following steps: establishing a BIM model of a target building; and based on the BIM model of the target building, point cloud extraction is carried out to determine first point cloud data of the target building; wherein the first point cloud data is point cloud data of the ground and the wall surface in the design geometric size; the construction site of the target building is scanned to obtain second point cloud data of the target building; wherein the second point cloud data is point cloud data of the ground and the wall surface in the field construction size; based on the first point cloud data and the second point cloud data, geometric precision comparison is carried out to determine the geometric deviation of the construction precision; the geometric deviation is standardized and exported to obtain a structured file; and the structured file is displayed through a visual interactive system. The above scheme can improve the precision and efficiency of the special-shaped wall geometric precision detection, and realize accurate monitoring and quality control of the construction process.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of the civil industry, in particular to a LiDAR-based construction precision control method and system for inclined keel partition walls. BACKGROUND

[0002] Building geometric quality is a key factor to ensure the safety of building structure, perfect function and aesthetic effect. Traditional quality detection methods, such as tape measure, total station, etc., are not only inefficient and labor-intensive, but also have discrete data and are easily affected by human factors, making it difficult to comprehensively and accurately evaluate complex structures.

[0003] Manual geometric precision detection of completed special-shaped walls is prone to safety hazards, and the detection method for geometric precision of special-shaped walls is low in precision and efficiency and not comprehensive. SUMMARY

[0004] The embodiments of the application aim to provide a LiDAR-based construction precision control method and system for inclined keel partition walls, which can improve the precision and efficiency of geometric precision detection of special-shaped walls and achieve accurate monitoring and quality control of the construction process.

[0005] The technical solution of the application is implemented as follows:

[0006] In a first aspect, the embodiments of the application provide a LiDAR-based construction precision control method for inclined keel partition walls, which comprises:

[0007] establishing a BIM model of a target building; and based on the BIM model of the target building, extracting point clouds to determine first point cloud data of the target building; wherein the first point cloud data is point cloud data of the ground and the wall surface in the design geometric size;

[0008] scanning the construction site of the target building to obtain second point cloud data of the target building; wherein the second point cloud data is point cloud data of the ground and the wall surface in the field construction size;

[0009] comparing geometric precision based on the first point cloud data and the second point cloud data to determine geometric deviation of construction precision; standardizing and exporting the geometric deviation to obtain a structured file; and displaying the structured file through a visual interactive system.

[0010] In the above solution, the extraction of point clouds based on the BIM model to determine the first point cloud data of the target building comprises:

[0011] based on the BIM model, performing point cloud format conversion on the target detection area in the target building to obtain first initial point cloud data;

[0012] data preprocessing is performed on the initial point cloud data to obtain processed first initial point cloud data;

[0013] Based on the processed first initial point cloud data, voxel thinning, normal vector estimation and double-threshold classification are performed to obtain the first point cloud data of the target building; wherein the first point cloud data contains color labels for distinguishing the wall and floor of the BIM model point cloud.

[0014] In the above scheme, the construction site of the target building is scanned to obtain the second point cloud data of the target building, comprising:

[0015] The construction site of the target building is scanned to obtain on-site scanning data;

[0016] Data preprocessing is performed on the on-site scanning data to obtain second initial point cloud data; and single-threshold segmentation is performed on the second initial point cloud data to obtain processed second initial point cloud data;

[0017] Based on the processed second initial point cloud data, voxel thinning, color space refinement and departure filtering noise reduction are performed to obtain the second point cloud data of the target building; wherein the first point cloud data contains color labels for distinguishing the wall and floor of the on-site point cloud.

[0018] In the above scheme, based on the first point cloud data and the second point cloud data, geometric precision comparison is performed to determine the geometric deviation of the construction precision, comprising:

[0019] Based on the first point cloud data and the second point cloud data, the respective dividing lines of the first point cloud data and the second point cloud data are determined; wherein the dividing line is the intersection line of the wall and the floor;

[0020] Based on the dividing line, color refinement separation and multi-threshold boundary detection are performed to determine the boundary points of the first point cloud data and the second point cloud data;

[0021] Based on the boundary points, reference point accurate positioning and multi-directional ray geometry detection are performed to determine the detection network of horizontal and vertical dimensions;

[0022] Based on the detection network, the error parameters of the construction precision are calculated to determine the geometric deviation of the construction precision.

[0023] In the above scheme, the dividing line includes a first dividing line and a second dividing line; the first dividing line corresponds to the first point cloud data; and the second dividing line corresponds to the second point cloud data;

[0024] Based on the boundary line, fine-grained color separation and multi-threshold boundary detection are performed to determine the boundary points of the first point cloud data and the second point cloud data, including:

[0025] Based on the first boundary line, fine-grained color matching is performed on the first point cloud data to determine the first color and the second color corresponding to the first point cloud data; wherein the first color and the second color are different.

[0026] Based on the second boundary line, fine-grained color matching is performed on the second point cloud data to determine the third and fourth colors corresponding to the second point cloud data; wherein, the first color, the second color, the third color, and the fourth color are all different;

[0027] For the first boundary line and the second boundary line, progressive threshold sequences are set respectively; and based on the progressive threshold sequences, boundary detection is performed to determine the initial boundary points of the first boundary line and the second boundary line respectively.

[0028] Based on the first color, the second color, the third color, and the fourth color, at least two target color points are selected respectively; based on the at least two target color points, the nearest neighbor distance is calculated; and based on the dual conditions and the nearest neighbor distance, the initial boundary points are filtered to obtain the filtered initial boundary points.

[0029] Outlier filtering and bidirectional cross-validation are performed on the initial boundary points after screening to determine the respective boundary points of the first point cloud data and the second point cloud data.

[0030] In the above scheme, the boundary points include a first boundary point and a second boundary point; wherein, the first boundary point corresponds to a first dividing line; the second boundary point corresponds to a second dividing line; and the detection network includes a first detection network and a second detection network.

[0031] The process of performing precise positioning of reference points and multi-directional ray geometric detection based on the boundary points to determine the detection network in the horizontal and vertical dimensions includes:

[0032] Based on the first point cloud data and the second point cloud data, the first global origin and the second global origin are determined respectively by the farthest point sampling algorithm;

[0033] Based on the first boundary line, traverse the first boundary points to determine the endpoint of the first reference direction; based on the second boundary line, traverse the second boundary points to determine the endpoint of the second reference direction.

[0034] Based on the first global origin and the first reference direction endpoint, a first vector is determined; and the first vector is subjected to horizontal point cloud feature detection and vertical spatial feature detection by horizontal rays and upward rays to determine the first detection network in horizontal and vertical dimensions.

[0035] Based on the second global origin and the second reference direction endpoint, a second vector is determined; and the second vector is subjected to horizontal point cloud feature detection and vertical spatial feature detection by horizontal rays and upward rays to determine the second detection network in the horizontal and vertical dimensions.

[0036] In the above scheme, the step of calculating the error parameters of the construction accuracy based on the detection network and determining the geometric deviation of the construction accuracy includes:

[0037] Based on the first and second detection networks in the detection network, the horizontal difference of the horizontal ray and the vertical difference of the vertical ray are calculated respectively; wherein, the horizontal difference characterizes the spatial deviation of different color boundaries in the horizontal direction between the design drawing and the actual construction; the vertical difference reflects the deviation characteristics in the vertical dimension;

[0038] Based on the horizontal and vertical differences, the distance deviation is determined;

[0039] Based on the first detection network, the first global origin, and the first reference direction endpoint, a first angle is calculated; based on the second detection network, the second global origin, and the second reference direction endpoint, a second angle is calculated; and based on the first angle and the second angle, the angle deviation is calculated.

[0040] The distance deviation and the angle deviation are analyzed and calculated to determine the error parameters of the construction accuracy, and the geometric deviation of the construction accuracy is determined based on the error parameters.

[0041] Secondly, embodiments of this application provide a LiDAR-based construction accuracy control system for inclined keel partition walls. This system includes: a system establishment module, a point cloud extraction module, an accuracy comparison module, and a display module.

[0042] The creation module is used to create a BIM model of the target building;

[0043] The point cloud extraction module is used to extract point clouds based on the BIM model of the target building to determine the first point cloud data of the target building; wherein, the first point cloud data is the point cloud data of the ground and walls with the design geometric dimensions; and to scan the construction site of the target building to obtain the second point cloud data of the target building; wherein, the second point cloud data is the point cloud data of the ground and walls with the on-site construction dimensions.

[0044] The accuracy comparison module is used to compare geometric accuracy based on the first point cloud data and the second point cloud data to determine the geometric deviation of the construction accuracy.

[0045] The display module is used to standardize and export the geometric deviation to obtain a structured file; and to display the structured file through a visual interactive system.

[0046] Thirdly, embodiments of this application provide a LiDAR-based device for controlling the construction accuracy of inclined keel partition walls, comprising: a processor and a memory; wherein,

[0047] The memory is used to store computer programs;

[0048] The processor is configured to call and run the computer program from the memory to perform the method as described in the first aspect.

[0049] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions for causing a processor to perform the method described in the first aspect.

[0050] This application provides a method and system for controlling the construction accuracy of inclined keel partition walls based on LiDAR. The method includes: establishing a BIM model of the target building; extracting point clouds based on the BIM model of the target building to determine the first point cloud data of the target building; wherein the first point cloud data is the point cloud data of the ground and wall surfaces with designed geometric dimensions; scanning the construction site of the target building to obtain the second point cloud data of the target building; wherein the second point cloud data is the point cloud data of the ground and wall surfaces with on-site construction dimensions; comparing the geometric accuracy based on the first point cloud data and the second point cloud data to determine the geometric deviation of the construction accuracy; standardizing and exporting the geometric deviation to obtain a structured file; and displaying the structured file through a visual interactive system. In the above solution, point cloud extraction is performed based on the BIM model of the target building to obtain the first point cloud data. The first point cloud data is then compared with the second point cloud data obtained by scanning to achieve intelligent detection of the geometric accuracy of irregular walls. Due to the use of intelligent detection, the accuracy and efficiency of the detection are higher than those of manual detection. In addition, it can also visualize and comprehensively reflect the geometric accuracy detection results between the actual values ​​on site and the design values ​​on the drawings. It can intuitively and quantitatively reveal the deviation between the design model and the actual construction results, thereby achieving precise monitoring and quality control of the construction process. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0052] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0053] Figure 1 An optional flowchart illustrating a method for controlling the construction accuracy of inclined keel partition walls based on LiDAR, provided for an embodiment of this application;

[0054] Figure 2 A system framework diagram of a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls provided in this application embodiment;

[0055] Figure 3 A BIM model point cloud of ground and wall color marking map for a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls provided in this application embodiment;

[0056] Figure 4 The on-site point cloud ground and wall color marking map of a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls provided in this application embodiment;

[0057] Figure 5 A BIM model point cloud map showing the boundary markers between the ground and wall areas for a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls, provided in this application embodiment;

[0058] Figure 6 A visual output diagram illustrating a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls, provided in an embodiment of this application;

[0059] Figure 7 A schematic diagram of a LiDAR-based inclined keel partition wall construction accuracy control system provided in this application embodiment;

[0060] Figure 8 This is a structural schematic diagram of a LiDAR-based inclined keel partition wall construction accuracy control device provided in an embodiment of this application. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0063] In the following description, references to "some embodiments," "this embodiment," "this application embodiment," and examples, etc., describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments and may be combined with each other without conflict.

[0064] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0065] This application provides a method for controlling the construction accuracy of inclined keel partition walls based on LiDAR. Figure 1 This is an optional flowchart illustrating a LiDAR-based method for controlling the construction accuracy of inclined keel partition walls, provided as an embodiment of this application. Figure 1 The steps shown are explained.

[0066] S101. Establish a BIM model of the target building; and based on the BIM model of the target building, extract point clouds to determine the first point cloud data of the target building; wherein, the first point cloud data is the point cloud data of the ground and walls with the design geometric dimensions.

[0067] In some embodiments of this application, Building Information Modeling (BIM) technology is currently the most advanced modeling method in the construction field with the highest degree of digitalization. This method can achieve information integration and sharing throughout the entire building lifecycle, improving the collaborative efficiency of various stages such as design, construction, and maintenance. BIM models contain rich design geometric and semantic information, enabling the digital representation of the physical and functional characteristics of target objects.

[0068] In some embodiments of this application, the LiDAR-based method for controlling the construction accuracy of inclined keel partition walls is adaptable to various wall inspection scenarios during construction.

[0069] In some embodiments of this application, the LiDAR-based method for controlling the construction accuracy of inclined keel partition walls is adapted to a LiDAR-based system for controlling the construction accuracy of inclined keel partition walls.

[0070] In some embodiments of this application, a detailed model of the target building is performed based on the design drawings to obtain a BIM model of the target building; the modeling accuracy directly affects the comparison of geometric dimensions. Based on the BIM model, the target detection area in the target building is converted into a point cloud format to obtain first initial point cloud data; the initial point cloud data is preprocessed to obtain processed first initial point cloud data; based on the processed first initial point cloud data, voxel thinning, normal vector estimation, and dual-threshold classification are performed to obtain the first point cloud data of the target building; wherein, the first point cloud data includes color markers to distinguish the walls and ground of the BIM model point cloud.

[0071] S102. Scan the construction site of the target building to obtain the second point cloud data of the target building; wherein, the second point cloud data is the point cloud data of the ground and walls at the construction site.

[0072] In some embodiments of this application, the construction site of the target building is scanned to obtain on-site scanning data; the on-site scanning data is preprocessed to obtain second initial point cloud data; and the second initial point cloud data is segmented by a single threshold to obtain processed second initial point cloud data; based on the processed second initial point cloud data, voxel thinning, color space refinement, and departure filtering noise reduction are performed to obtain the second point cloud data of the target building; wherein, the first point cloud data includes color markers to distinguish the walls and ground of the on-site point cloud.

[0073] S103. Based on the first point cloud data and the second point cloud data, compare the geometric accuracy to determine the geometric deviation of the construction accuracy; export the geometric deviation in a standardized manner to obtain a structured file; and display the structured file through a visual interactive system.

[0074] In some embodiments of this application, based on first point cloud data and second point cloud data, the respective boundary lines of the first point cloud data and second point cloud data are determined; wherein, the boundary line is the intersection line of the wall and the ground; based on the boundary lines, fine color separation and multi-threshold boundary detection are performed to determine the respective boundary points of the first point cloud data and second point cloud data; based on the boundary points, precise positioning of reference points and multi-directional ray geometric detection are performed to determine the detection network in the horizontal and vertical dimensions; based on the detection network, error parameters of construction accuracy are calculated to determine the geometric deviation of construction accuracy. The geometric deviation is standardized and exported to obtain a structured file; and the structured file is displayed through a visual interactive system.

[0075] For example, Figure 2 The framework of the LiDAR-based inclined keel partition wall construction accuracy control system is shown, which includes the following three main parts: (1) BIM design geometric dimension point cloud extraction; (2) on-site (Scan) construction dimension point cloud extraction; (3) geometric accuracy result output and visualization.

[0076] Understandably, by extracting point cloud data from the BIM model of the target building, the first point cloud data is obtained. This first point cloud data is then compared with the second point cloud data obtained through scanning, enabling intelligent detection of the geometric accuracy of irregular walls. Due to the use of intelligent detection, the accuracy and efficiency of the detection are higher than those of manual detection. In addition, it can visualize and comprehensively reflect the geometric accuracy detection results between the actual values ​​on site and the design values ​​on the drawings. It can intuitively and quantitatively reveal the deviation between the design model and the actual construction results, thereby achieving precise monitoring and quality control of the construction process.

[0077] In some embodiments of this application, the point cloud extraction based on the BIM model in step S101 to determine the first point cloud data of the target building can be implemented through steps S201-S203, as follows:

[0078] S201. Based on the BIM model, the target detection area in the target building is converted into point cloud format to obtain the first initial point cloud data.

[0079] S202. Perform data preprocessing on the initial point cloud data to obtain the processed first initial point cloud data.

[0080] S203. Based on the processed first initial point cloud data, voxel thinning, normal vector estimation, and dual threshold classification are performed to obtain the first point cloud data of the target building; wherein, the first point cloud data includes color markers to distinguish the walls and ground of the BIM model point cloud.

[0081] In some embodiments of this application, voxel thinning refers to adaptive voxel thinning when the number of points in the first initial point cloud data after preprocessing is greater than 50,000, and it is only used for normal vector estimation; if the density is less than 1,000 points, it is passed as is to prevent loss of details.

[0082] For example, firstly, there's the processing of the BIM model at the design level. A detailed model of the target building needs to be created based on the design drawings. The modeling accuracy directly affects the comparison of geometric dimensions. After establishing the BIM model of the target building, the target detection area is converted into a point cloud format to obtain the first initial point cloud data. Secondly, a binary color-coded base map is constructed for direct use by the boundary extraction module. The process consists of three steps: removing NaN / out-of-limit coordinates and calculating residuals; parallel computing of bounding boxes to provide scale references for subsequent downsampling. For lightweight computation, adaptive voxel thinning is used when the number of points exceeds 50,000, only for normal vector estimation; points with a density of less than 1,000 are passed as is to prevent loss of detail. Normals are estimated using a KD-Tree, and coarsely classified into "plan / elevation" categories based on the absolute value of the Z component using dual thresholds. The output is a color-coded point cloud, with geometric coordinates corresponding to each input point. At this point, the obtained point cloud file has the ground and walls marked by color, and the first point cloud data is as follows: Figure 3 As shown, orange represents the walls and red represents the floor.

[0083] Understandably, obtaining the first point cloud data of the design drawings through the BIM model can provide a basis for comparing the subsequent design with the on-site construction.

[0084] In some embodiments of this application, S102 can be implemented by S301-S303, as follows:

[0085] S301. Scan the construction site of the target building to obtain on-site scanning data.

[0086] S302. Perform data preprocessing on the on-site scanning data to obtain second initial point cloud data; and perform single threshold segmentation on the second initial point cloud data to obtain processed second initial point cloud data.

[0087] S303. Based on the processed second initial point cloud data, voxel thinning, color space refinement, and departure filtering noise reduction are performed to obtain the second point cloud data of the target building; wherein, the first point cloud data contains color markers to distinguish the walls and ground of the site point cloud.

[0088] For example, the actual point cloud data for target detection is obtained from the construction site. Since construction sites are generally complex, the acquired data needs to be preprocessed after exporting to obtain a second initial point cloud data. In the original coordinate system, binary semantic segmentation is performed using a single threshold for the Z-component of the normal vector, coarsely dividing the wall points into "plane" and "elevation" categories, each assigned a different base color. For large-scale data, voxel-random hierarchical downsampling is implemented to compress the computational load. Then, color space refinement is introduced, using hue-saturation dual-limiting to remove unwanted colors, forcibly normalizing to the base color, and optionally employing statistical outlier filtering to further suppress isolated noise points. The final output second point cloud data can be directly used for registration with the BIM model and subsequent boundary extraction, meeting the requirements for high-precision construction verification. The second point cloud data is as follows: Figure 4 As shown, yellow represents the walls and green represents the floor.

[0089] Understandably, obtaining actual construction data on-site through 3D laser scanning replaces the traditional manual climbing and measurement for inspection, reducing the risks to personnel.

[0090] In some embodiments of this application, S103 can be implemented by S401-S404, as follows:

[0091] S401. Based on the first point cloud data and the second point cloud data, determine the boundary lines of the first point cloud data and the second point cloud data respectively; wherein, the boundary line is the intersection line of the wall and the ground.

[0092] In some embodiments of this application, the boundary line includes a first boundary line and a second boundary line; the first boundary line corresponds to the first point cloud data; and the second boundary line corresponds to the second point cloud data.

[0093] In some embodiments of this application, the first point cloud data and the second point cloud data are preprocessed to determine the respective boundary lines of the first point cloud data and the second point cloud data.

[0094] S402. Based on the boundary line, perform fine color separation and multi-threshold boundary detection to determine the boundary points of the first point cloud data and the second point cloud data respectively.

[0095] In some embodiments of this application, the boundary points include a first boundary point and a second boundary point; wherein the first boundary point corresponds to a first dividing line; the second boundary point corresponds to a second dividing line; and the detection network includes a first detection network and a second detection network.

[0096] In some embodiments of this application, based on a first boundary line, fine-grained color matching is performed on the first point cloud data to determine a first color and a second color corresponding to the first point cloud data; wherein the first color and the second color are different. Based on a second boundary line, fine-grained color matching is performed on the second point cloud data to determine a third color and a fourth color corresponding to the second point cloud data; wherein the first color, the second color, the third color, and the fourth color are all different. For the first boundary line and the second boundary line, progressive threshold sequences are set respectively. Based on the progressive threshold sequences, boundary detection is performed to determine the initial boundary points of the first boundary line and the second boundary line respectively. Based on the first color, the second color, the third color, and the fourth color, at least two target color points are selected respectively. Based on the at least two target color points, the nearest neighbor distance is calculated. Based on the dual conditions and the nearest neighbor distance, the initial boundary points are filtered to obtain the filtered initial boundary points. Outlier filtering and bidirectional cross-validation are performed on the filtered initial boundary points to determine the boundary points of the first point cloud data and the second point cloud data respectively.

[0097] For example, fine-grained color separation refers to assigning different colors to point clouds at different locations in the first and second point cloud data respectively. The BIM point cloud (i.e., the first point cloud data) prioritizes accurate matching of pure colors, and then performs secondary verification using multiple restrictions of R, G, and B, as well as the max-min difference (for example, red requires R≥220, G≤60, B≤60 and RG>150, orange requires R≥200, G∈[100,180], B≤80 and RG>50, with a classification accuracy of ≥98%). The on-site point cloud (i.e., the second point cloud data) first locks in the pure color yellow (255, 255, 0), then uses conditions such as GR>50 and GB>100 to distinguish green, and uses |RG|<80 and B<120 to lock in yellow to prevent yellow-green confusion). To avoid large point clouds slowing down the process, the sampling step size is dynamically adjusted to 2 when BIM > 80,000 or on-site > 400,000; otherwise, the step size is kept at 1. The four types of point counts are output throughout the process. If color detection is abnormal, a quality warning is immediately triggered, prompting the user to check the parameters.

[0098] It should be noted that the first color is red, the second color is orange, the third color is yellow, and the fourth color is green.

[0099] Multi-threshold progressive boundary detection involves setting different thresholds for boundary detection. First, a KD-tree nearest neighbor search framework is constructed to ensure the efficiency and reliability of boundary point retrieval. For the target boundary, based on the target color point cloud, a KD-tree is constructed, setting the number of nearest neighbors to 3 to ensure the spatial correlation of boundary points. Simultaneously, the entire KD-tree initialization and search process is implemented, with real-time anomaly capture and error log output to prevent program crashes due to data anomalies.

[0100] For the wall and ground boundaries of the BIM point cloud, progressive thresholds of 0.08f, 0.15f, 0.25f, 0.35f, and 0.45f are set to adapt to different accuracy requirements. For the wall and ground boundaries of the on-site point cloud, a threshold sequence of 0.12f, 0.22f, 0.32f, and 0.42f is set to better match the boundary features of the on-site point cloud. During detection, for each source color point (i.e., the first, second, third, and fourth colors), three target color points are searched, and the nearest neighbor distance (dist) is calculated. Only points that satisfy "0.005f < dist < current threshold" and "the difference between the nearest and second nearest neighbor distances < threshold × 50%" are retained as boundary points. Isolated points and noise points are excluded through dual conditions to ensure boundary accuracy.

[0101] To further improve boundary accuracy, an IQR outlier filtering mechanism is introduced: collect the nearest neighbor distances of all initially detected boundary points, sort them in ascending order, calculate the median (median_dist), 25th percentile (Q1), and 75th percentile (Q3), determine the outlier range using the interquartile range (IQR = Q3 - Q1), filter outliers with a distance < Q1 - 1.5 × IQR or > Q3 + 1.5 × IQR, output the number of removed points, and improve the accuracy of boundary point extraction.

[0102] Finally, bidirectional cross-validation is used to refine boundary integrity: forward validation searches for target color points (e.g., orange) from source color points (e.g., red) to obtain the initial boundary; backward validation constructs a KD-tree of the source color point cloud, searches for source color points (e.g., red) from the target color points (e.g., orange), and uses the median value of the threshold sequence (e.g., 0.25f for white boundaries) as the backward threshold to supplement boundary points missed by forward detection; finally, forward and backward boundary points are merged, and deduplication is performed using an index to improve boundary coverage and ensure that there are no significant missing boundaries. Figure 5 As shown, the white point cloud represents the boundary points of the first point cloud data.

[0103] S403. Based on boundary points, perform precise positioning of reference points and multi-directional ray geometric detection to determine the detection network in the horizontal and vertical dimensions.

[0104] In some embodiments of this application, based on first point cloud data and second point cloud data, a first global origin and a second global origin are determined respectively using a farthest point sampling algorithm; based on a first boundary line, the first boundary points are traversed to determine a first reference direction endpoint; based on a second boundary line, the second boundary points are traversed to determine a second reference direction endpoint; based on the first global origin and the first reference direction endpoint, a first vector is determined; and horizontal point cloud feature detection and vertical spatial feature detection are performed on the first vector using horizontal rays and upward rays to determine a first detection network for horizontal and vertical dimensions; based on the second global origin and the second reference direction endpoint, a second vector is determined; and horizontal point cloud feature detection and vertical spatial feature detection are performed on the second vector using horizontal rays and upward rays to determine a second detection network for horizontal and vertical dimensions.

[0105] For example, precise positioning of the reference point is the spatial reference basis for ray detection and deviation quantification. The core revolves around the precise positioning of the global origin (point A) and the endpoint of the reference direction (point B). Through multiple strategies and anomaly handling, the integrity of subsequent ray coverage and the reliability of direction are ensured.

[0106] It should be noted that the global origin (point A) includes both the first and second global origins, meaning that a global origin is included in different point cloud datasets. The reference direction endpoint (point B) includes both the first and second reference direction endpoints, meaning that a reference direction endpoint is included in different point cloud datasets.

[0107] In global origin (point A) localization, the farthest point sampling (FPS) algorithm is used to accurately capture the geometric center of the point cloud. The sampling strategy is dynamically adjusted according to the point cloud size: when the number of points in the point cloud is ≤500, the Euclidean distance of all point pairs is calculated, and the point pair with the largest distance is directly selected; when the number of points is >500, uniform sampling is performed with a step size of "total number of points ÷ 500", which reduces the amount of computation while preserving global features and avoids inefficiency caused by exhaustive calculation. To cope with extreme scenarios where the point cloud is too dense and the effective farthest point pair cannot be identified, the two points with the smallest and largest X-axis coordinates are selected as the farthest point pair, and the midpoint between the two is taken as point A to ensure that the localization does not fail. After localization, the XYZ coordinates of point A and the distance of the farthest point pair are output. Geometric verification ensures that point A is in the central region of the point cloud space, so that the subsequent ray coverage is ≥90%, providing a basis for full-area deviation detection.

[0108] The endpoint of the reference direction (point B) is located using the white boundary as the core reference. By traversing the boundary point cloud, the point with the largest X-axis coordinate is selected as point B. This point can accurately reflect the spatial extension direction of the white boundary and provide a clear endpoint for the ray reference direction. Then, a validity check of the B point coordinates is performed. By judging whether the coordinate is NaN or an infinite value, invalid point interference is eliminated. If the coordinate is invalid, the program is terminated directly to avoid the entire ray detection being deviated due to the incorrect reference direction.

[0109] Multi-directional ray geometry detection constructs a detection network covering horizontal and vertical dimensions through scientific ray direction generation and precise intersection point search, providing data support for subsequent distance and angle deviation calculations.

[0110] In the ray direction generation mechanism, the core reference direction is first determined based on reference points A and B: the AB vector (obtained by subtracting the coordinates of point A from the coordinates of point B) is calculated, normalized, and then projected onto the XOY plane. Simultaneously, the angle between this projected vector and the positive X-axis (reference angle) is calculated, serving as the initial reference for the ray direction. To achieve full spatial coverage, 71 direction vectors are generated by expanding in 5° increments: including one reference direction (denoted as AB1), 35 clockwise directions (AB2-AB36), and 35 counterclockwise directions (AB37-AB71). All direction vectors are normalized to ensure consistent length references during subsequent ray emission, avoiding detection errors caused by differences in vector scale.

[0111] The ray intersection search logic is designed with a dual dimension of "horizontal (all horizontal is relative to the ground) + vertical", corresponding to the AB / AC series and AD / AE series rays respectively;

[0112] Horizontal Rays (AB / AC Series): Focus on point cloud feature detection in the horizontal direction. AB series rays are emitted horizontally from point A, searching the boundary point cloud to obtain target intersections. During the search, a step size of 0.2 and a maximum detection distance of 2000 are set, with an initial search radius of 5. If no intersection is captured, the radius is increased to 10 to improve the success rate. AC series rays use the same search logic. After each ray detection is completed, the intersection distance, endpoint coordinates, and whether a substitute color marker is enabled are recorded. A successful intersection is counted in `successCount`, and a failed intersection is counted in `failCount`. Finally, the ray success rate is calculated and output by "success count ÷ total number of rays", quantifying the effectiveness of horizontal direction detection.

[0113] Upward Rays (AD / AE Series): Targets spatial feature detection in the vertical dimension. For each horizontal direction, upward rays with elevation angles ranging from 5° to 90° are generated at 5° step sizes (18 angles in total). The direction vector is calculated as "horizontal direction vector × cos(elevation angle) + Z-axis unit vector × sin(elevation angle)," and normalized before being used for searching. The AD series uses fixed-search BIM point clouds for design, while the AE series uses fixed-search actual site point clouds. Search parameters are set to a step size of 0.5, a maximum distance of 1000, and a radius of 3 to ensure vertical detection accuracy. After finding the intersection point, the distance data is directly output to provide a vertical dimension reference for subsequent angle deviation calculations.

[0114] It should be noted that an upward ray is a vertical ray.

[0115] S404. Based on the detection network, calculate the error parameters of construction accuracy and determine the geometric deviation of construction accuracy.

[0116] In some embodiments of this application, based on a first detection network and a second detection network in the detection network, the horizontal difference of the horizontal ray and the vertical difference of the vertical ray are calculated respectively; wherein, the horizontal difference characterizes the spatial deviation of different color boundaries in the horizontal direction between the drawing design and the actual construction; the vertical difference reflects the deviation characteristics in the vertical dimension; the distance deviation is determined based on the horizontal difference and the vertical difference; a first angle is calculated based on the first detection network, the first global origin, and the endpoint of the first reference direction; a second angle is calculated based on the second detection network, the second global origin, and the endpoint of the second reference direction; and the angle deviation is calculated based on the first angle and the second angle; the distance deviation and the angle deviation are analyzed and calculated to determine the error parameters of the construction accuracy, and the geometric deviation of the construction accuracy is determined according to the error parameters.

[0117] For example, in distance deviation calculation, it is carried out separately in the horizontal and vertical dimensions: the horizontal deviation focuses on the AB and AC series rays, traverses the ray data with the same index of the two, and calculates the difference of "AB ray length - AC ray length", which reflects the spatial deviation of different color boundaries in the horizontal direction between the drawing design and the actual construction; the vertical deviation is for the AD and AE series rays, and the rays with the same spatial direction are matched by the combination index of "basic index + upward index", and the difference of "AD ray length - AE ray length" is calculated to accurately reflect the deviation characteristics in the vertical dimension.

[0118] The angle deviation calculation uses point A as the spatial reference to construct an angle quantification model: When calculating the angle ∠ABD, A is the vertex, and points B (the endpoint of ray AB) and D (the endpoint of ray AD) are the edge points. First, the vectors AB and AD are obtained by the difference between the coordinates of the two points. Then, the dot product formula "cosθ = AB" is used. .The cosine of the included angle is calculated using AD / (|AB|×|AD|), and converted to degrees. The calculation logic for ∠ACE is similar, using A as the vertex, C (the endpoint of ray AC), and E (the endpoint of ray AE) as edge points, and the angle is obtained through the dot product of vectors AC and AE. The angle difference is calculated by matching ∠ABD and ∠ACE at the same spatial position using a combination of "basic index - upward index," calculating the difference between "∠ACE - ∠ABD," recording whether there is any substitution color influence, and quantifying the angular deviation within the vertical plane.

[0119] In the statistical parameter calculation stage, in-depth analysis is performed on all distance and angle difference indicators: the average value is calculated to reflect the overall deviation trend; an average value close to 0 indicates a small and relatively balanced deviation. The standard deviation is calculated to measure the degree of data dispersion; a smaller standard deviation indicates a more stable deviation and more controllable engineering accuracy. The maximum and minimum values ​​are extracted to clarify the deviation range, providing a basis for locating extreme deviation positions. Strict precision format control is maintained during output; distance parameters are retained to 4 decimal places, and angle parameters are retained to 2 decimal places, ensuring data traceability and comparability.

[0120] For standardized export of CSV files, the ExcelExporter module generates more than ten types of structured files: 00_Summary Report.csv serves as an index file, containing descriptions of the purpose of all files and methods for parsing combined indexes; 01_Basic Information.csv records basic point cloud data (total number of points, number of valid points, number of points of each color), ray detection efficiency (success rate, substitution rate), and A / B reference point coordinates, providing support for data overview; 02-05_Distance Data.csv is categorized by AB / AC / AD / AE series, recording in detail the index, label, distance, endpoint coordinates, and whether clockwise / using a substitute color for each ray; 06-07_Difference Data.csv, 08-09_Angle Data.csv, and 10_Angle Difference.csv store specific data and substitution instructions for distance difference, angle value, and angle difference, respectively; 11_Statistical Analysis.csv summarizes the average, standard deviation, extreme values, and other statistical parameters of all deviation indicators; 12_Line Segment Data.csv fully records the start / end point, length, color, and substitution marker for each ray segment. The error parameters are exported as a CSV file and standardized for output. Table 1 shows an example of a partial output of the angle difference.

[0121] Table 1

[0122]

[0123] The 3D visualization interactive system presents point clouds and deviation features in an intuitive way: the original point cloud is loaded semi-transparently with a point size of 1, preserving the overall outline without obscuring key markers; for example... Figure 6As shown, the reference elements are highlighted—point A is marked with a red sphere (radius 5.0), point B with a blue sphere (radius 4.0), the AB reference line is drawn in green (line width 5), and the line connecting the farthest points is drawn in orange (line width 3), enabling quick location of the spatial reference; ray segments are distinguished by type (AB primary color is blue, AB substitution is orange, AC primary color is red, AC substitution is yellow, AD is cyan, AE is magenta), with a line width of 1, providing clear visual distinction; the optimized marking design further enhances readability—AB series end Add 3D text markers (primary color: dark blue sphere + white text; alternative color: red sphere + white text with "*") to each index. Add a gold sphere (radius 3.0) + yellow text to every 10 indices to highlight key positions. Interactive operations support left-click rotation, right-click panning, and scroll wheel zooming. Use the keyboard "r" key to reset the view and "q" key to exit. Set the initial camera position to diagonally behind point A (X-200, Y-200, Z+100) to ensure the global point cloud and ray network are visible, facilitating engineers' intuitive observation of deviation distribution.

[0124] Understandably, in terms of safety, using 3D laser scanning to obtain actual on-site construction data replaces traditional manual climbing measurements, reducing the risks to personnel. Through multi-dimensional deviation calculations (distance, angle) and visualization, issues such as excessive deviations between wall verticality and design values ​​are identified, preventing structural safety hazards caused by geometrical discrepancies. In terms of efficiency, multi-directional X-ray inspection replaces traditional manual measurement with total station measurements, improving efficiency and reducing labor costs. Multi-directional X-ray inspection is applicable to various wall surfaces and is relatively universal. In terms of cost, through accuracy comparison, geometric deviations are detected mid-construction (rather than after completion), avoiding large-scale rework (such as wall demolition and floor re-leveling), reducing redundant investment in materials and labor. In terms of results management, the results are intuitively visualized; the 3D visualization system clearly presents the deviation distribution, supports interactive operation, and allows the construction team to quickly locate areas requiring rectification. Statistical parameters (mean, standard deviation) help management assess the overall accuracy level, providing data reference for subsequent project optimization.

[0125] Based on the LiDAR-based method for controlling the construction accuracy of inclined keel partition walls described in the above embodiments, this application also provides a LiDAR-based system for controlling the construction accuracy of inclined keel partition walls, such as... Figure 7 As shown, Figure 7 This application provides a schematic diagram of a LiDAR-based inclined keel partition wall construction accuracy control system. The LiDAR-based inclined keel partition wall construction accuracy control system 7 includes: a system establishment module 701, a point cloud extraction module 702, an accuracy comparison module 703, and a display module 704.

[0126] The establishment module 701 is used to establish a BIM model of the target building;

[0127] The point cloud extraction module 702 is used to extract point clouds based on the BIM model of the target building to determine the first point cloud data of the target building; wherein, the first point cloud data is the point cloud data of the ground and walls with the design geometric dimensions; and to scan the construction site of the target building to obtain the second point cloud data of the target building; wherein, the second point cloud data is the point cloud data of the ground and walls with the on-site construction dimensions.

[0128] The accuracy comparison module 703 is used to compare the geometric accuracy based on the first point cloud data and the second point cloud data to determine the geometric deviation of the construction accuracy.

[0129] The display module 704 is used to standardize and export the geometric deviation to obtain a structured file; and to display the structured file through a visual interactive system.

[0130] Based on the LiDAR-based method for controlling the construction accuracy of inclined keel partition walls described in the above embodiments, this application also provides a LiDAR-based device for controlling the construction accuracy of inclined keel partition walls, such as... Figure 8 As shown, Figure 8 This is a schematic diagram of a LiDAR-based inclined keel partition wall construction accuracy control device provided in an embodiment of this application. The LiDAR-based inclined keel partition wall construction accuracy control device 8 includes a processor 801 and a memory 802. The memory 802 is used to store computer programs; the processor 801 is used to call and run the computer programs from the memory to execute the LiDAR-based inclined keel partition wall construction accuracy control method as described in the above embodiment.

[0131] In the embodiments of this application, the processor 801 described above can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field-Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that for different devices, the electronic device used to implement the above processor function can also be other types, and the embodiments of this application do not specifically limit it.

[0132] This application provides a computer-readable storage medium storing a computer program for implementing, when executed by a processor, the LiDAR-based method for controlling the construction accuracy of inclined keel partition walls as described in any of the above embodiments.

[0133] For example, the program instructions corresponding to the LiDAR-based inclined keel partition wall construction accuracy control method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to the LiDAR-based inclined keel partition wall construction accuracy control method in the storage media are read or executed by an electronic device, the LiDAR-based inclined keel partition wall construction accuracy control method as described in any of the above embodiments can be realized.

[0134] Furthermore, in the embodiments of this application, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0135] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0136] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, these will not be repeated here.

[0137] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0138] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0139] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0140] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0141] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0142] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0143] The above description is merely an embodiment of this application, but the protection scope of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for controlling the construction accuracy of inclined keel partition walls based on LiDAR, characterized in that, The method includes: A BIM model of the target building is established; and based on the BIM model of the target building, point cloud extraction is performed to determine the first point cloud data of the target building; wherein, the first point cloud data is the point cloud data of the ground and walls with the design geometric dimensions; The construction site of the target building is scanned to obtain the second point cloud data of the target building; wherein, the second point cloud data is the point cloud data of the ground and walls at the construction site. Based on the first point cloud data and the second point cloud data, a geometric accuracy comparison is performed to determine the geometric deviation of the construction accuracy; the geometric deviation is standardized and exported to obtain a structured file; and the structured file is displayed through a visual interactive system. The step of comparing geometric accuracy based on the first point cloud data and the second point cloud data to determine the geometric deviation of construction accuracy includes: Based on the first point cloud data and the second point cloud data, the boundary lines of the first point cloud data and the second point cloud data are determined respectively; wherein, the boundary line is the intersection line of the wall and the ground; the boundary line includes a first boundary line and a second boundary line; the first boundary line corresponds to the first point cloud data; the second boundary line corresponds to the second point cloud data; Based on the boundary line, fine-grained color separation and multi-threshold boundary detection are performed to determine the boundary points of the first point cloud data and the second point cloud data, including: based on the first boundary line, fine-grained color matching is performed on the first point cloud data to determine the first color and the second color corresponding to the first point cloud data; wherein the first color and the second color are different; based on the second boundary line, fine-grained color matching is performed on the second point cloud data to determine the third color and the fourth color corresponding to the second point cloud data; wherein the first color, the second color, the third color, and the fourth color are all different; for the first boundary line and the... The second boundary line is defined by setting progressive threshold sequences. Boundary detection is performed based on these progressive threshold sequences to determine the initial boundary points for both the first and second boundary lines. At least two target color points are selected based on the first, second, third, and fourth colors. The nearest neighbor distance is calculated based on these at least two target color points. The initial boundary points are then filtered based on a combination of conditions and the nearest neighbor distance to obtain filtered initial boundary points. Outlier filtering and bidirectional cross-validation are performed on the filtered initial boundary points to determine the boundary points for both the first and second point cloud data. Based on the boundary points, precise positioning of reference points and multi-directional ray geometric detection are performed to determine the detection network in the horizontal and vertical dimensions. Based on the detection network, the error parameters of the construction accuracy are calculated, and the geometric deviation of the construction accuracy is determined.

2. The method according to claim 1, characterized in that, The step of extracting point clouds based on the BIM model to determine the first point cloud data of the target building includes: Based on the BIM model, the target detection area in the target building is converted into point cloud format to obtain the first initial point cloud data; The initial point cloud data is preprocessed to obtain the processed first initial point cloud data; Based on the processed first initial point cloud data, voxel thinning, normal vector estimation, and dual threshold classification are performed to obtain the first point cloud data of the target building; wherein, the first point cloud data includes color markers to distinguish the walls and ground of the BIM model point cloud.

3. The method according to claim 1, characterized in that, The step of scanning the construction site of the target building to obtain the second point cloud data of the target building includes: The construction site of the target building is scanned to obtain on-site scanning data; The on-site scanning data is preprocessed to obtain second initial point cloud data; and the second initial point cloud data is segmented by a single threshold to obtain processed second initial point cloud data. Based on the processed second initial point cloud data, voxel thinning, color space refinement, and departure filtering noise reduction are performed to obtain the second point cloud data of the target building; wherein, the first point cloud data includes color markers to distinguish the walls and ground of the site point cloud.

4. The method according to claim 1, characterized in that, The boundary points include a first boundary point and a second boundary point; wherein the first boundary point corresponds to a first dividing line; the second boundary point corresponds to a second dividing line; the detection network includes a first detection network and a second detection network; The process of performing precise positioning of reference points and multi-directional ray geometric detection based on the boundary points to determine the detection network in the horizontal and vertical dimensions includes: Based on the first point cloud data and the second point cloud data, the first global origin and the second global origin are determined respectively by the farthest point sampling algorithm; Based on the first boundary line, traverse the first boundary points to determine the endpoint of the first reference direction; based on the second boundary line, traverse the second boundary points to determine the endpoint of the second reference direction. Based on the first global origin and the first reference direction endpoint, a first vector is determined; and the first vector is subjected to horizontal point cloud feature detection and vertical spatial feature detection by horizontal rays and upward rays to determine the first detection network in horizontal and vertical dimensions. Based on the second global origin and the second reference direction endpoint, a second vector is determined; and the second vector is subjected to horizontal point cloud feature detection and vertical spatial feature detection by horizontal rays and upward rays to determine the second detection network in the horizontal and vertical dimensions.

5. The method according to claim 1, characterized in that, The step of calculating the error parameters of the construction accuracy based on the detection network and determining the geometric deviation of the construction accuracy includes: Based on the first and second detection networks in the detection network, the horizontal difference of the horizontal ray and the vertical difference of the vertical ray are calculated respectively; wherein, the horizontal difference characterizes the spatial deviation of different color boundaries in the horizontal direction between the design drawing and the actual construction; the vertical difference reflects the deviation characteristics in the vertical dimension. Based on the horizontal and vertical differences, the distance deviation is determined; Based on the first detection network, the first global origin, and the first reference direction endpoint, a first angle is calculated; based on the second detection network, the second global origin, and the second reference direction endpoint, a second angle is calculated; and based on the first angle and the second angle, the angle deviation is calculated. The distance deviation and the angle deviation are analyzed and calculated to determine the error parameters of the construction accuracy, and the geometric deviation of the construction accuracy is determined based on the error parameters.

6. A LiDAR-based system for controlling the construction accuracy of inclined keel partition walls, characterized in that, The LiDAR-based inclined keel partition wall construction accuracy control system includes: a modeling module, a point cloud extraction module, an accuracy comparison module, and a display module. The creation module is used to create a BIM model of the target building; The point cloud extraction module is used to extract point clouds based on the BIM model of the target building to determine the first point cloud data of the target building; wherein, the first point cloud data is the point cloud data of the ground and walls with the design geometric dimensions; and to scan the construction site of the target building to obtain the second point cloud data of the target building; wherein, the second point cloud data is the point cloud data of the ground and walls with the on-site construction dimensions. The accuracy comparison module is used to compare the geometric accuracy based on the first point cloud data and the second point cloud data to determine the geometric deviation of the construction accuracy. The accuracy comparison module is further configured to determine the respective boundary lines of the first point cloud data and the second point cloud data based on the first point cloud data and the second point cloud data; wherein the boundary line is the intersection line of the wall and the ground; the boundary line includes a first boundary line and a second boundary line; the first boundary line corresponds to the first point cloud data; the second boundary line corresponds to the second point cloud data; based on the boundary lines, perform fine color separation and multi-threshold boundary detection to determine the respective boundary points of the first point cloud data and the second point cloud data; based on the boundary points, perform precise positioning of reference points and multi-directional ray geometric detection to determine the detection network in the horizontal and vertical dimensions; based on the detection network, calculate the error parameters of the construction accuracy to determine the geometric deviation of the construction accuracy; The precision comparison module is further configured to perform fine-grained color matching on the first point cloud data based on the first boundary line to determine the first color and the second color corresponding to the first point cloud data; wherein the first color and the second color are different; perform fine-grained color matching on the second point cloud data based on the second boundary line to determine the third color and the fourth color corresponding to the second point cloud data; wherein the first color, the second color, the third color, and the fourth color are all different; set progressive threshold sequences for the first boundary line and the second boundary line respectively; and perform boundary detection based on the progressive threshold sequences to determine the initial boundary points of the first boundary line and the second boundary line respectively; select at least two target color points based on the first color, the second color, the third color, and the fourth color respectively; calculate the nearest neighbor distance based on the at least two target color points; and filter the initial boundary points based on the dual conditions and the nearest neighbor distance to obtain the filtered initial boundary points; and perform outlier filtering and bidirectional cross-validation on the filtered initial boundary points to determine the boundary points of the first point cloud data and the second point cloud data respectively. The display module is used to standardize and export the geometric deviation to obtain a structured file; and to display the structured file through a visual interactive system.

7. A LiDAR-based device for controlling the construction accuracy of inclined keel partition walls, characterized in that, include: Processor and memory, of which, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores executable instructions for causing a processor to execute, thereby implementing the method of any one of claims 1 to 5.

Images