Civil engineering cost management method based on artificial intelligence

Abstract

The invention discloses a civil engineering cost management method based on artificial intelligence, and relates to civil engineering cost. The method comprises the following steps: constructing a BIM design model of a construction project; a laser radar is adopted to collect point cloud data of a construction site, and an actual construction point cloud model is generated; performing spatial registration on the actual construction point cloud model and the BIM design model to obtain a registered fusion model; according to the fusion model, a completed component is identified through a deep learning algorithm, and the actual completed work amount is obtained through volume calculation; comparing the actual completed project amount with a planned completed project amount of a corresponding time node in the BIM design model, and calculating a progress deviation rate; and adjusting the construction cost according to the progress deviation rate in combination with the unit price information in the bill of quantities. Aiming at the low precision of civil engineering cost dynamic monitoring, the laser point cloud and the BIM model are fused, the deep learning algorithm is utilized to identify the completion state of the component and calculate the actual engineering amount, and the dynamic monitoring precision of the engineering cost is improved.

Description

Technical Field

[0001] The present application relates to the field of engineering cost management, and in particular to a civil engineering cost management method based on artificial intelligence. Background Art

[0002] Civil engineering projects are characterized by large investment scales, long construction cycles, and numerous stakeholders. Project cost management runs through the entire project lifecycle and is a key link in ensuring the project's economic benefits. With the construction industry's transition to digitalization and intelligentization, traditional cost management models can no longer meet the needs of refined management of modern engineering projects. Especially during the construction phase, accurately understanding the project's completion status in real time and adjusting cost plans in a timely manner are of great significance for controlling project costs and avoiding investment risks. The application of Building Information Modeling (BIM) technology provides a new technical means for project cost management. By constructing a three-dimensional digital model containing geometric information, material properties, and cost data, it enables the automatic extraction of project quantities and the accurate calculation of costs. However, BIM models reflect design intent and planned status. Comparing and analyzing the actual construction status with the BIM model to achieve dynamic cost management based on actual completion remains a technical challenge facing the industry.

[0003] Traditional methods rely on manual on-site measurement using tools such as tape measures and levels, which is not only time-consuming and labor-intensive, but also has measurement accuracy that is greatly affected by human factors. For high-altitude operations, complex structures, and other difficult-to-reach areas, measurement work presents safety hazards. Manual measurements usually use a sampling method, which makes it difficult to cover all components, resulting in omissions in engineering quantity statistics. Existing methods often use estimation to determine the completed engineering quantity, such as roughly estimating the completion percentage based on the completion status of each floor, or inferring the completed quantity based on material consumption. This method ignores the actual completion form of the component, especially for partially completed components, whose completed volume cannot be accurately calculated. Fixed coefficients are often used for steel bar deductions and masonry mortar joint calculations of reinforced concrete components, which deviate from the actual situation.

[0004] The collection and aggregation of project progress data is typically performed on a monthly basis, resulting in significant lags. By the time progress deviations are discovered, the optimal opportunity for adjustment has often been missed. Insufficient correlation analysis between sub-projects makes it difficult to promptly identify progress risks affecting the critical path. Progress information is disconnected from cost data, making it impossible to achieve dynamic cost forecasting based on actual progress. Summary of the Invention

[0005] In response to the low accuracy of dynamic monitoring of civil engineering costs, this application provides a civil engineering cost management method based on artificial intelligence. By integrating laser point cloud with BIM model, deep learning algorithm is used to identify the completion status of components and calculate the actual engineering quantity, etc., thereby improving the dynamic monitoring accuracy of engineering costs.

[0006] The present application provides a civil engineering cost management method based on artificial intelligence, including: S1, constructing a BIM design model of a construction project, the BIM design model including component information, a bill of quantities, and planned progress data of the construction phase; S2, using a laser radar to collect point cloud data of the construction site to generate an actual construction point cloud model; S3, spatially aligning the actual construction point cloud model with the BIM design model to obtain a fusion model after alignment; S4, based on the fusion model, identifying completed components through a deep learning algorithm, and obtaining the actual completed project quantity through volume calculation; S5, comparing the actual completed project quantity with the planned completed project quantity at the corresponding time node in the BIM design model, and calculating the progress deviation rate; S6, adjusting the project cost based on the progress deviation rate in combination with the unit price information in the bill of quantities.

[0007] Furthermore, S4, based on the fusion model, identifies completed components through a deep learning algorithm, and obtains the actual completed engineering volume through volume calculation, including: extracting point cloud feature data from the fusion model, the point cloud feature data including spatial coordinates, reflection intensity and normal vector; voxelizing the point cloud feature data according to the spatial coordinates, and converting the continuous point cloud data into a three-dimensional voxel grid, each voxel unit contains the reflection intensity and normal vector information of the corresponding spatial position; inputting the three-dimensional voxel grid into a pre-trained convolutional neural network model to obtain component type labels, spatial positions and component point cloud clusters; according to the component type labels and spatial position information, matching the identified component point cloud clusters with standard components of corresponding positions and types in the BIM design model, and determining the completion status of the component by calculating the point cloud coverage, the completion status including fully completed, partially completed and not started.

[0008] When the point cloud coverage is greater than the preset threshold Q1, the component is determined to be completely completed. The system obtains the unique identifier of the component in the BIM model through the component matching results; the design volume properties of the component are directly queried through the data interface of the BIM model; the design volume value obtained by the query is directly assigned to the actual completed volume; this processing method avoids calculation errors in point cloud reconstruction and improves efficiency and accuracy.

[0009] When the point cloud coverage is less than the preset threshold Q1 and greater than the preset threshold Q2, it is determined to be a partially completed component. According to the spatial coordinates of the point cloud cluster of the corresponding component, the three-dimensional surface of the component point cloud cluster is reconstructed, and the volume contained in the closed surface is calculated. The calculated volume value is used as the actual completed volume of the component; according to the actual completed volume of each component and the volume calculation rules of the corresponding component type in the bill of quantities, the actual completed project volume of each sub-item is accumulated according to the sub-items of the project to obtain the actual completed project volume of each sub-item.

[0010] In particular, for fully completed components R>Q1, this application avoids cumulative errors in the point cloud data processing process by directly calling the precise design volume in the BIM model. When the component is basically completed, the deviation between the actual geometry and the designed geometry is within the allowable range of the project. Directly using the design value not only ensures accuracy, but also greatly improves calculation efficiency. For partially completed components Q2≤R≤Q1, the volume of the actual completed part is accurately calculated through point cloud reconstruction technology, which solves the technical problem that traditional methods cannot accurately evaluate components under construction. For unstarted components R<Q2, it avoids misjudging scattered noise points as the start of construction, thereby improving the robustness of the system.

[0011] Furthermore, spatial coordinates are used to determine the spatial position and geometric shape of components; reflection intensity is used to distinguish components of different materials; and normal vectors are used to identify the surface orientation and boundaries of components.

[0012] Furthermore, the three-dimensional voxel grid is input into the pre-trained convolutional neural network model to obtain the component type label, spatial position and component point cloud cluster, including: the convolutional neural network model includes a feature extraction layer, an instance segmentation layer and a semantic recognition layer; the feature extraction layer uses reflection intensity and normal vector information to extract the geometric features and material features of the component; the instance segmentation layer separates adjacent components through a region growing algorithm based on the extracted geometric features, and outputs point cloud clusters of each independent component; the semantic recognition layer outputs the component type label corresponding to each point cloud cluster based on the extracted material features and geometric features; according to the boundary voxel coordinates of the component point cloud cluster, the center position of each component is calculated as the spatial position.

[0013] Furthermore, based on the component type label and spatial location information, the identified component point cloud clusters are matched with standard components of corresponding positions and types in the BIM design model. The completion status of the component is determined by calculating the point cloud coverage, including:

[0014] Based on the component type label, a set of candidate components of the same type is screened in the BIM design model; the centroid coordinates of the component point cloud cluster are calculated as the center position C. Among them, n is the number of points in the component point cloud cluster, (x i ,y i ,z i ) is the three-dimensional coordinate of the i-th point; calculate the distance D between the center position C and the geometric center of each component in the candidate component set, Select the component with the smallest distance D as the matching component; obtain the axis-aligned bounding box AABB parameters of the matching component, the AABB parameters include the minimum vertex coordinates (x min ,y min ,z min ) and the maximum vertex coordinate (x max ,y max ,zmax ); Compare the coordinates of each point in the component point cloud cluster with the axis-aligned bounding box AABB boundary, and count the number of points N that fall into the axis-aligned bounding box AABB in ; Divide the axis-aligned bounding box AABB into M×N×K voxel units according to the preset resolution r, count the number of points in each voxel unit, and calculate the actual distribution density

[0015]

[0016] According to the geometric information of the matching components in the BIM design model, the theoretical point cloud density ρ is calculated by simulating the laser radar scanning process. standard ; Calculate the theoretical number of sampling points per unit volume based on the laser radar scanning parameters and the surface area of ​​the component; Calculate the point cloud coverage When the actual distribution density is detected to be significantly uneven, the point cloud is divided into multiple connected areas through the connected domain analysis algorithm, and the number of points in the largest connected domain is calculated as N. in The coverage ratio ε is modified to: R' = R × ε. When R > Q1, the process is considered fully completed; when Q2 ≤ R ≤ Q1, the process is considered partially completed; and when R < Q2, the process is considered not started. Q1 and Q2 are preset coverage ratio thresholds. The value range of Q1 is 90% to 100%; the value range of Q2 is 10% to 20%.

[0017] Specifically, component type labels identified through deep learning reduce the matching search space from thousands of components in the full model to dozens of components of the same type, reducing search complexity from O(n) to O(n / k), where k is the number of component types. This introduction of semantic prior knowledge avoids mismatches that can occur with purely geometric matching, where components are similar in form but different in nature.

[0018] Second, the center of mass As the geometric feature center of the point cloud cluster, it has translation invariance and noise robustness. Performing a nearest neighbor search ensures that the components closest in spatial location are correctly matched. This centroid-based matching method is computationally efficient, with a complexity of only O(m), where m is the number of candidate components.

[0019] Finally, ρ is calculated by simulating the lidar scanning process standard This method takes into account the impact of practical factors such as scanning angle, distance, and occlusion on point cloud density. This theoretical value calculation based on a physical model is closer to reality than the simple uniform distribution assumption and solves the coverage consistency problem under different scanning conditions.

[0020] Furthermore, when the point cloud coverage is less than a preset threshold Q1 and greater than a preset threshold Q2, the component is determined to be partially completed. Based on the spatial coordinates of the corresponding component's point cloud cluster, the three-dimensional surface of the component's point cloud cluster is reconstructed, and the volume contained in the closed surface is calculated. The calculated volume value is used as the actual completed volume of the component. This includes: outlier processing of the partially completed component's point cloud cluster to obtain a valid point cloud; extracting the matching component's construction direction information from the BIM design model as the main construction direction vector. Civil engineering construction has clear directional characteristics: columns: vertical construction from bottom to top; beams: horizontal construction along the long axis; slabs: layered casting from bottom to top. The component's type attributes and geometric parameters are read from the BIM model; the main construction direction vector is automatically determined based on the component type; and for special components, custom directions can be read from the BIM model's construction information attributes.

[0021] The three-dimensional point cloud data is projected one-dimensionally along the main construction direction vector, simplifying the spatial distribution problem into a linear distribution problem. The projection value of each point along the construction direction represents its position in the construction process. The range of the projection value is determined, that is, the distance from the construction starting point to the current farthest point. This range is divided into m sub-ranges (the m value is adaptively adjusted according to the component size, generally 30-50). The number of points in each sub-range is counted and divided by the range length to obtain the linear density value. This forms a density distribution curve along the construction direction.

[0022] Furthermore, at the construction interface, the point cloud for the completed portion is dense, while the point cloud for the unfinished portion is sparse or absent, resulting in a sharp drop in density at the interface. Starting from the construction start point, the density ratio of adjacent subintervals is calculated one by one. When the density ratio at a certain location falls below a threshold value T (empirical value 0.3), it indicates a sudden change in density; this sudden change is the completed construction interface.

[0023] The cutting plane passes through the identified construction interface points, with the plane normal serving as the primary construction direction vector, ensuring that the cutting plane divides the component into completed and unfinished parts. The BIM software's built-in geometry engine (such as ACIS and Parasolid) is used to perform an intersection operation between the solid and half-space, retaining the geometry on the construction start side. The cut geometry is a standard B-Rep solid model, and the geometry engine's volume calculation function is directly invoked. The resulting volume value is a precise mathematical calculation result without cumulative error, and is used as the actual completed volume of the partially completed component.

[0024] In particular, this application transforms the traditional point cloud surface reconstruction problem into a geometric cutting problem based on the BIM model. Traditional point cloud surface reconstruction, such as the AlphaShape algorithm, needs to infer the complete geometric shape from discrete point clouds, which is a reverse engineering "from part to whole", and is bound to have information missing and reconstruction errors. However, this solution directly uses the precise geometry provided by the BIM model to simplify the problem to "determine the cutting position", which is a forward analysis "from the whole to the part", and in principle avoids the uncertainty of geometric reconstruction. In addition, the selection of the α parameter of the AlphaShape algorithm is subjective. Different α values ​​will produce different boundary shapes, and there is no guarantee that the reconstructed boundary is consistent with the actual construction interface. This solution directly locates the construction interface through density analysis, and then constructs a standard cutting plane to reduce the three-dimensional boundary recognition problem to a one-dimensional interface detection problem, which in principle guarantees the uniqueness and accuracy of the boundary.

[0025] Furthermore, according to the actual completed volume of each component and the volume calculation rules of the corresponding component type in the bill of quantities, the actual completed volume of each sub-project is obtained by accumulating the sub-items of the project, including: for reinforced concrete components, the calculation is based on the following rules: beam and column components: actual completed volume = actual completed volume × (1-0.025), where 0.025 is the steel volume deduction coefficient; plate components: actual completed volume = actual completed volume × (1-0.015), where 0.015 is the steel volume deduction coefficient; foundation components: actual completed volume = actual completed volume, without deducting the steel volume; for steel structure components, the calculation is based on the following rules: actual completed volume = actual completed volume × 7850kg / m 3 ÷1000.

[0026] Furthermore, the volume calculation rules also include: for masonry components, the calculation is based on the following rules: standard brick masonry: actual completed project volume = actual completed volume × 0.95, where 0.95 is the reduction coefficient for considering mortar joints; block masonry: actual completed project volume = actual completed volume ÷ (block volume × 1.1), to obtain the number of blocks, where 1.1 is the mortar joint coefficient; for formwork engineering, the calculation is based on the following rules: beam formwork: actual completed project volume = (beam bottom width + 2 × beam height) × actual completed length; slab formwork: actual completed project volume = actual completed area; all component project volumes under the same sub-project code are added up according to the same measurement unit to obtain the actual completed project volume of the sub-item.

[0027] Furthermore, S5 compares the actual completed project volume with the planned completed project volume at the corresponding time node in the BIM design model, and calculates the progress deviation rate, including: obtaining the current construction date, extracting the planned completed project volume at the corresponding time node from the planned progress data of the BIM design model; comparing the actual completed project volume of each sub-project with the corresponding planned completed project volume, and calculating the progress deviation rate of each sub-project: progress deviation rate = (actual completed project volume - planned completed project volume) / planned completed project volume × 100%; according to the weight coefficient of each sub-project in the total project , calculate the weighted average progress deviation rate, where the weight coefficient is determined according to the proportion of the contract amount of each sub-project; when the progress deviation rate of a sub-project exceeds the preset warning threshold, generate progress warning information, which includes the name of the sub-project, the deviation rate value and the analysis of the cause of the deviation; judge the project progress status according to the positive and negative values ​​of the progress deviation rate: when the progress deviation rate is positive, it means that the project progress is ahead of schedule; when the progress deviation rate is negative, it means that the project progress is lagging behind; generate a progress deviation analysis report, which includes the planned project quantity, actual project quantity, deviation rate and cumulative completion percentage of each sub-project.

[0028] Compared with the existing technology, the advantages of this application are:

[0029] This application uses high-density point cloud data (with a resolution of millimeters) collected by lidar to accurately align and fuse with the BIM design model, and uses a three-dimensional convolutional neural network to automatically identify components and calculate point cloud coverage, thereby achieving accurate calculation of engineering quantities.

[0030] For fully completed components (coverage R>90%), the precise design volume in the BIM model is directly extracted to avoid point cloud reconstruction errors; for partially completed components (10%≤R≤90%), the construction interface is accurately identified through density gradient analysis, and the BIM geometry engine is used to perform Boolean cutting operations to obtain the precise volume.

[0031] Compared to traditional manual measurement and estimation methods, this method reduces the error in engineering quantity calculation from 10% to 15% to 2% to 3%. This method, in particular for complex, irregularly shaped components and hard-to-reach high-altitude locations, enables full coverage and high-precision automated measurement. Furthermore, refined engineering quantity calculation rules have been established for different material types, such as differentiated deduction coefficients for reinforced concrete components (2.5% for beams and columns, 1.5% for slabs) and precise conversion of masonry mortar joints. This ensures accurate conversion from actual completed volume to engineering quantity, providing a reliable data foundation for dynamic cost monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present application will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same numbers represent the same structures, wherein:

[0033] Figure 1 is an exemplary flow chart of a civil engineering cost management method based on artificial intelligence according to some embodiments of the present application;

[0034] Figure 2 This is an exemplary flow chart for obtaining the actual amount of completed work according to some embodiments of the present application;

[0035] Figure 3 is an exemplary flow chart of generating a component point cloud cluster according to some embodiments of the present application;

[0036] Figure 4 This is an exemplary flowchart of calculating point cloud coverage according to some embodiments of the present application. DETAILED DESCRIPTION

[0037] The method and system provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, S1, builds a BIM design model of the construction project, which includes component information, bill of quantities and planned progress data in the construction phase; S2, uses lidar to collect point cloud data of the construction site to generate an actual construction point cloud model; S3, spatially aligns the actual construction point cloud model with the BIM design model to obtain a fusion model after alignment; S4, based on the fusion model, identifies completed components through a deep learning algorithm, and obtains the actual completed project quantity through volume calculation; S5, compares the actual completed project quantity with the planned completed project quantity at the corresponding time node in the BIM design model, and calculates the progress deviation rate; S6, adjusts the project cost based on the progress deviation rate and the unit price information in the bill of quantities.

[0039] Specifically, S1 constructs a BIM design model of the construction project, which includes component information, bill of quantities, and planned progress data during the construction phase.

[0040] S2 uses LiDAR to collect point cloud data of the construction site and generate an actual construction point cloud model; uses a LiDAR scanner to perform a 360-degree full-scale scan of the construction site at preset time intervals; pre-processes the collected raw point cloud data, including denoising, alignment, and segmentation; uses a deep learning-based semantic segmentation algorithm to classify the point cloud data and identify different materials and component types; and generates an actual construction point cloud model with semantic labels.

[0041] S3, spatially aligning the actual construction point cloud model with the BIM design model to obtain a fusion model after alignment, including: extracting the characteristic geometric elements of the building from the BIM design model, including wall corner points, column center points and floor boundary lines, to generate a BIM feature point set; detecting plane features from the actual construction point cloud model using the RANSAC algorithm, identifying wall surfaces, floor surfaces and column surfaces, and extracting plane intersection lines and intersection points as a point cloud feature point set; performing a preliminary match between the BIM feature point set and the point cloud feature point set, and using the nearest point iterative algorithm to calculate the initial transformation matrix T0, including the rotation matrix R0 and the translation vector t0; performing a coarse alignment on the actual construction point cloud model based on the initial transformation matrix T0, and converting the point cloud coordinate system to the BIM coordinate system; based on the coarse alignment, using the point-to-surface ICP algorithm for fine alignment: projecting each point in the point cloud to the nearest component surface in the BIM model; calculating the point-to-surface distance, and constructing an error function Among them, d i is the distance from the i-th point to the corresponding surface; the transformation parameters are iteratively optimized by the least squares method until the error function converges; the final transformation matrix T is applied final The actual construction point cloud model is transformed to obtain a registered point cloud in the same coordinate system as the BIM design model; the registered point cloud data is integrated with the geometric data of the BIM design model to generate a fusion model containing design information and actual construction information.

[0042] like Figure 2 As shown, S4, based on the fusion model, identifies the completed components through deep learning algorithms and obtains the actual completed engineering volume through volume calculation; extracts point cloud feature data from the fusion model, which includes spatial coordinates, reflection intensity and normal vector; spatial coordinates are used to determine the spatial position and geometric shape of the component; reflection intensity is used to distinguish components of different materials; normal vector is used to identify the surface orientation and boundary of the component. i The spatial coordinates of i ,y i ,z i ), where i∈[1,N], N is the total number of point clouds; reflection intensity Among them, P r is the received power, P e is the emission power, the reflection intensity characteristic value of different materials: Concrete: I i ∈[0.3,0.5], steel: I i ∈[0.7,0.9], brickwork: I i ∈[0.4,0.6], the reflection intensity is clustered by Gaussian mixture model (GMM) to realize automatic material recognition.

[0043] Calculate the normal vector and use the principal component analysis (PCA) algorithm to calculate the normal vector n of each point i =(n x ,n y ,n z ), select point P i The k-nearest neighbor point set N k (P i ), k = 20, construct the covariance matrix Among them, p j ∈N k (P i ), normal vector n i is the eigenvector corresponding to the minimum eigenvalue of C. When the angle θ between the normal vectors of adjacent points is greater than 30°, it is determined to be a boundary.

[0044] The point cloud feature data is voxelized according to the spatial coordinates, and the continuous point cloud data is converted into a three-dimensional voxel grid. Each voxel unit contains the reflection intensity and normal vector information of the corresponding spatial position. Specifically, the voxel resolution v is set r = 0.05m (5cm), calculate the point cloud bounding box: [x min ,x max ]×[y min ,y max ]×[z min ,z max ]. Generate voxel grid size:

[0045] The three-dimensional voxel grid is input into a pre-trained convolutional neural network model to obtain component type labels, spatial positions, and component point cloud clusters. The convolutional neural network model includes a feature extraction layer, an instance segmentation layer, and a semantic recognition layer. The feature extraction layer uses reflection intensity and normal vector information to extract the geometric and material features of the component. The instance segmentation layer separates adjacent components based on the extracted geometric features using a region growing algorithm and outputs point cloud clusters for each independent component. The semantic recognition layer outputs the component type label corresponding to each point cloud cluster based on the extracted material and geometric features. The center position of each component is calculated as the spatial position based on the boundary voxel coordinates of the component point cloud cluster.

[0046] like Figure 3 and Figure 4 As shown in the figure, according to the component type label and spatial position information, the identified component point cloud clusters are matched with the standard components of the corresponding position and type in the BIM design model. The completion status of the component is determined by calculating the point cloud coverage, and the completion status includes fully completed, partially completed and not started. Based on the component type label, the candidate component set of the same type is screened in the BIM design model. The centroid coordinates of the component point cloud cluster are calculated as the center position C. Among them, n is the number of points in the component point cloud cluster, (x i ,y i ,z i ) is the three-dimensional coordinate of the i-th point; calculate the distance D between the center position C and the geometric center of each component in the candidate component set,

[0047] Select the component with the smallest distance D as the matching component; obtain the axis-aligned bounding box AABB parameters of the matching component, the AABB parameters include the minimum vertex coordinates (x min ,y min ,z min ) and the maximum vertex coordinate (x max ,y max ,z max ); Compare the coordinates of each point in the component point cloud cluster with the axis-aligned bounding box AABB boundary, and count the number of points N that fall into the axis-aligned bounding box AABB in ; Divide the axis-aligned bounding box AABB into M×N×K voxel units according to the preset resolution r, count the number of points in each voxel unit, and calculate the actual distribution density According to the geometric information of the matching components in the BIM design model, the theoretical point cloud density ρ is calculated by simulating the laser radar scanning process. standard ; Calculate the theoretical number of sampling points per unit volume based on the laser radar scanning parameters and the surface area of ​​the component; Calculate the point cloud coverage When significant unevenness in the actual distribution density is detected, the point cloud is divided into multiple connected regions using a connected domain analysis algorithm. The proportion ε of the number of points in the largest connected domain to Nin is calculated, and the coverage rate is corrected: R' = R × ε. When R > Q1, the process is considered fully completed; when Q2 ≤ R ≤ Q1, it is considered partially completed; and when R < Q2, it is considered not yet started. Q1 and Q2 are preset coverage rate thresholds. The value range of Q1 is 90% to 100%; the value range of Q2 is 10% to 20%.

[0048] When the point cloud coverage R>Q1 (Q1=0.95), the component is determined to be completely completed, and the unique identifier of the component is set as CID (ComponentID); the design volume is extracted through the BIM data interface function: V actual =BIM.GetVolume(CID), where V actual Indicates the actual completed volume, in m 3 , the designed volume is directly used as the actual completed volume to avoid point cloud reconstruction errors.

[0049] When Q2≤R≤Q1(Q2=0.3,Q1=0.95), the refinement process is executed.

[0050] First, remove outliers, for the point cloud cluster P = {p1,p2,.....,p n Each point p in i , calculate p i The average distance to its k nearest neighbors: Where k = 20 (empirical value); calculate the statistical characteristics of the average distance of all points: Outlier determination criteria: If Then p i For outliers, remove all outliers and get the valid point cloud P valid .

[0051] Extract the main construction direction vector v from the BIM model according to the component type T: If T = "column": v = (0,0,1) T ; If T = "beam": If T = "plate": v = (0,0,1) T ; If T = "wall": v = n wall (wall normal vector); where p start ,p end are the starting and ending coordinates of the beam respectively.

[0052] For the valid point cloud P valid Every point p in i =(x i ,y i ,z i ) T Projection: t i =p i ×v=x i v x +y i v y +z i v z ; Determine the projection interval: [t min ,t max ];t min =min{t i |i=1,2,.....,n};t max =max{t i |i=1,2,.....,n}.

[0053] Divide the projection interval into m subintervals (m=50): The jth subinterval: I j =[t min +(j-1)Δt,t min +jΔt], j=1,2,.....,m. Count the number of points in each subinterval: n j =|{p i|t i ∈I j}|, calculate the line density: (Unit: points / meter).

[0054] Identify the construction interface location and calculate the density ratio of adjacent sub-intervals: Construction interface determination: If r j <T(T=0.3) and ρ j+1 <ρ min , then the interface is located at the end of the jth subinterval, where ρ min =0.1×max{ρ j |j=1,2,.....,m} is the minimum density threshold, construction interface position: h cut =t min +j×Δt.

[0055] Cut and calculate the partially completed volume, cutting point: P cut =p origin +h cut ×v, where p origin Is the starting point of the component. Cutting plane equation: v×(PP cut )=0, expanded form: v x (xx cut )+v y (yy cut )+v z (zz cut )=0.

[0056] Boolean cutting operation: define half space H: H = {P|v×(PP cut )≤0}. Perform the intersection operation:

[0057] G partial =G BIM ∩H, where G BIM is the component geometry in BIM, G partial For the cut part. Volume calculation: V actual =∫∫∫ Gpartial dV actually gets its precise value through the Volume() function of the geometry engine.

[0058] According to the actual completed volume of each component and the volume calculation rules of the corresponding component type in the bill of quantities, the actual completed volume of each sub-project is obtained by adding up the sub-items of the project.

[0059] For reinforced concrete components, the calculation is based on the following rules: Beam and column components: actual completed project quantity = actual completed volume × (1-0.025), where 0.025 is the steel bar volume deduction coefficient; Plate components: actual completed project quantity = actual completed volume × (1-0.015), where 0.015 is the steel bar volume deduction coefficient; Foundation components: actual completed project quantity = actual completed volume, without deducting the steel bar volume.

[0060] For steel structure components, the calculation is as follows: Actual completed project volume = actual completed volume × 7850kg / m 3 ÷1000.

[0061] For masonry components, the calculation is based on the following rules: Standard brick masonry: Actual completed project volume = Actual completed volume × 0.95, where 0.95 is the reduction factor for considering mortar joints; Block masonry: Actual completed project volume = Actual completed volume ÷ (Block volume × 1.1), to obtain the number of blocks, where 1.1 is the mortar joint coefficient.

[0062] For formwork engineering, the calculation is based on the following rules: Beam formwork: Actual completed engineering quantity = (beam bottom width + 2 × beam height) × actual completed length; Plate formwork: Actual completed engineering quantity = actual completed area. The engineering quantities of all components under the same sub-item code are accumulated using the same measurement unit to obtain the actual completed engineering quantity for that sub-item.

[0063] S5, comparing the actual completed project volume with the planned completed project volume at the corresponding time node in the BIM design model, and calculating the progress deviation rate; obtaining the current construction date, and extracting the planned completed project volume at the corresponding time node from the planned progress data of the BIM design model;

[0064] Compare the actual completed work volume of each sub-project with the corresponding planned completed work volume, and calculate the progress deviation rate of each sub-project: progress deviation rate = (actual completed work volume - planned completed work volume) / planned completed work volume × 100%;

[0065] Calculate the weighted average schedule deviation rate based on the weight coefficient of each sub-project in the overall project, where the weight coefficient is determined based on the proportion of the contract amount of each sub-project; when the schedule deviation rate of a sub-project exceeds the preset warning threshold, generate a schedule warning message, which includes the sub-project name, deviation rate value, and deviation cause analysis;

[0066] The project progress status is determined based on the positive or negative value of the progress deviation rate: when the progress deviation rate is positive, it indicates that the project is ahead of schedule; when the progress deviation rate is negative, it indicates that the project is behind schedule; a progress deviation analysis report is generated, which includes the planned project quantity, actual project quantity, deviation rate and cumulative completion percentage of each sub-project.

[0067] S6, adjust the project cost based on the progress deviation rate and the unit price information in the bill of quantities.

[0068] The invention of the present application and its implementation methods are described schematically above. This description is not restrictive. Without departing from the spirit or basic features of the present application, the present application can be implemented in other specific forms. What is shown in the accompanying drawings is only one of the implementation methods of the invention of the present application, and the actual structure is not limited to this. Therefore, if a person of ordinary skill in the art is inspired by it, without departing from the purpose of the invention, a structural method and embodiment similar to the technical solution are designed without creativity, which should all fall within the scope of protection of the present application. In addition, the word "including" does not exclude other elements or steps, and the word "one" before an element does not exclude the inclusion of "multiple" elements. Words such as first and second are used to indicate names and do not indicate any specific order.

Claims

1. A civil engineering cost management method based on artificial intelligence, characterized in that: include: S1, build a BIM design model for the construction project, which includes component information, bill of quantities, and planned progress data for the construction phase; S2, uses LiDAR to collect point cloud data of the construction site and generate a point cloud model of the actual construction; S3, spatially registering the actual construction point cloud model with the BIM design model to obtain a registered fusion model; S4, based on the fusion model, identifies completed components through deep learning algorithms and calculates the actual completed project volume; S5, compare the actual completed project volume with the planned completed project volume at the corresponding time node in the BIM design model, and calculate the progress deviation rate; S6, adjust the project cost based on the progress deviation rate and the unit price information in the bill of quantities.

2. The civil engineering cost management method based on artificial intelligence according to claim 1 is characterized in that: S4, based on the fusion model, uses a deep learning algorithm to identify completed components and calculates the actual completed project volume through volume calculation, including: Extracting point cloud feature data from the fusion model, where the point cloud feature data includes spatial coordinates, reflection intensity, and normal vector; The point cloud feature data is voxelized according to the spatial coordinates, and the continuous point cloud data is converted into a three-dimensional voxel grid. Each voxel unit contains the reflection intensity and normal vector information of the corresponding spatial position; The 3D voxel grid is input into a pre-trained convolutional neural network model to obtain component type labels, spatial locations, and component point cloud clusters. Based on the component type label and spatial location information, the identified component point cloud clusters are matched with the standard components of the corresponding position and type in the BIM design model. The completion status of the component is determined by calculating the point cloud coverage. The completion status includes fully completed, partially completed, and not started. When the point cloud coverage is greater than the preset threshold Q1, the component is determined to be fully completed, and the design volume of the corresponding component is directly extracted from the BIM design model as the actual completed volume; When the point cloud coverage is less than the preset threshold Q1 and greater than the preset threshold Q2, it is determined to be a partially completed component. Based on the spatial coordinates of the point cloud cluster of the corresponding component, the three-dimensional surface of the component point cloud cluster is reconstructed, and the volume contained in the closed surface is calculated. The calculated volume value is used as the actual completed volume of the component; According to the actual completed volume of each component and the volume calculation rules of the corresponding component type in the bill of quantities, the actual completed volume of each sub-project is obtained by adding up the sub-items of the project.

3. The civil engineering cost management method based on artificial intelligence according to claim 2 is characterized in that: Spatial coordinates are used to determine the spatial position and geometric shape of components; Reflection intensity is used to distinguish components of different materials; Normal vectors are used to identify the surface orientation and boundaries of components.

4. The civil engineering cost management method based on artificial intelligence according to claim 2 is characterized in that: The 3D voxel grid is fed into a pre-trained convolutional neural network model to obtain component type labels, spatial locations, and component point cloud clusters, including: The convolutional neural network model includes a feature extraction layer, an instance segmentation layer, and a semantic recognition layer; The feature extraction layer uses reflection intensity and normal vector information to extract the geometric features and material features of the component; The instance segmentation layer separates adjacent components through the region growing algorithm based on the extracted geometric features and outputs point cloud clusters of each independent component; The semantic recognition layer outputs the component type label corresponding to each point cloud cluster based on the extracted material features and geometric features; According to the boundary voxel coordinates of the component point cloud cluster, the center position of each component is calculated as the spatial position.

5. The civil engineering cost management method based on artificial intelligence according to claim 2 is characterized in that: Based on the component type label and spatial location information, the identified component point cloud clusters are matched with the standard components of the corresponding position and type in the BIM design model. The completion status of the component is determined by calculating the point cloud coverage, including: Based on component type labels, filter candidate component sets of the same type in the BIM design model; Calculate the centroid coordinates of the component point cloud cluster as the center position C, Among them, n is the number of points in the component point cloud cluster, (x i ,y i ,z i ) is the three-dimensional coordinate of the i-th point; Calculate the distance D between the center position C and the geometric center of each component in the candidate component set, Select the component with the smallest distance D as the matching component; Get the axis-aligned bounding box AABB parameters of the matching component, which include the minimum vertex coordinates (x min ,y min ,z min ) and the maximum vertex coordinate (x max ,y max ,z max ); Compare the coordinates of each point in the component point cloud cluster with the axis-aligned bounding box AABB boundary, and count the number of points N that fall into the axis-aligned bounding box AABB in ; Divide the axis-aligned bounding box AABB into M×N×K voxel units according to the preset resolution r, count the number of points in each voxel unit, and calculate the actual distribution density According to the geometric information of the matching components in the BIM design model, the theoretical point cloud density ρ is calculated by simulating the laser radar scanning process. standard ; Calculate the theoretical number of sampling points per unit volume based on the laser radar scanning parameters and the surface area of ​​the component; Calculate the point cloud coverage R= r actual ×100%, when the actual distribution density is detected to be obviously uneven, the standard The connected domain analysis algorithm divides the point cloud into multiple connected regions and calculates the number of points in the largest connected domain, N. in The ratio ε is used to correct the coverage rate: R'=R×ε; When R>Q1, it is determined to be fully completed; when Q2≤R≤Q1, it is determined to be partially completed; when R<Q2, it is determined to be not started, where Q1 and Q2 are preset coverage thresholds.

6. The civil engineering cost management method based on artificial intelligence according to claim 5 is characterized by: The value range of Q1 is 90% to 100%; The value range of Q2 is 10% to 20%.

7. The civil engineering cost management method based on artificial intelligence according to claim 4 is characterized in that: When the point cloud coverage is less than the preset threshold Q1 and greater than the preset threshold Q2, the component is determined to be partially completed. The three-dimensional surface of the component point cloud cluster is reconstructed according to the spatial coordinates of the corresponding component point cloud cluster, and the volume contained in the closed surface is calculated. The calculated volume value is used as the actual completed volume of the component, including: Perform outlier processing on the point cloud cluster of the partially completed component to obtain a valid point cloud; Extract the construction direction information of the matching components from the BIM design model as the main construction direction vector; Project the valid point cloud along the main construction direction vector, divide the projection interval into m subintervals, count the number of points in each subinterval, and calculate the density distribution of the valid point cloud in the main construction direction; Calculate the density ratio of adjacent subintervals. r i When the value is less than the preset threshold T, the end of the i-th subinterval ρ i+1 The location is determined as the construction interface location; According to the construction interface position, a cutting plane is set on the geometric model of the matching component in the BIM design model. The cutting plane passes through the construction interface position and is perpendicular to the main construction direction vector. The geometry engine in BIM is used to perform Boolean cutting operations on the geometric models of the matching components, retaining the part between the construction start end and the cutting plane, and calculating the volume of the retained part as the actual completed volume of the partially completed component.

8. The artificial intelligence-based civil engineering cost management method according to claim 7, characterized in that: Based on the actual completed volume of each component and the volume calculation rules of the corresponding component type in the bill of quantities, the actual completed volume of each sub-project is obtained by accumulating the sub-items of the project, including: For reinforced concrete components, the calculation is based on the following rules: Beam and column components: Actual completed project quantity = actual completed volume × (1-0.025), where 0.025 is the steel bar volume deduction factor; Plate components: Actual completed project volume = actual completed volume × (1-0.015), where 0.015 is the steel bar volume deduction factor; Foundation components: actual completed project volume = actual completed volume, excluding steel bar volume; For steel structural members, calculations are carried out according to the following rules: Actual completed project volume = actual completed volume × 7850kg / m 3 ÷1000.

9. The civil engineering cost management method based on artificial intelligence according to claim 8 is characterized in that: Volume calculation rules, including: For masonry components, the calculation is based on the following rules: Standard brickwork: actual completed work volume = actual completed volume × 0.95, where 0.95 is the reduction factor for considering mortar joints; Block masonry: Actual completed work volume = actual completed volume ÷ (block volume × 1.1), to get the number of blocks, where 1.1 is the mortar joint coefficient; For template projects, calculations are made according to the following rules: Beam formwork: actual completed work quantity = (beam bottom width + 2 × beam height) × actual completed length; Plate formwork: actual completed project quantity = actual completed area; Add up the quantities of all components under the same sub-project code using the same unit of measurement to obtain the actual completed quantity of the sub-project.

10. The civil engineering cost management method based on artificial intelligence according to any one of claims 2 to 9, characterized in that: S5: Compare the actual completed project volume with the planned completed project volume at the corresponding time node in the BIM design model and calculate the progress deviation rate, including: Obtain the current construction date and extract the planned completion amount of the corresponding time node from the planned progress data of the BIM design model; Compare the actual completed work volume of each sub-project with the corresponding planned completed work volume, and calculate the progress deviation rate of each sub-project: Schedule deviation rate = (actual completed work volume - planned completed work volume) / planned completed work volume × 100%; Calculate the weighted average schedule deviation rate based on the weight coefficient of each sub-project in the overall project, where the weight coefficient is determined based on the proportion of the contract amount of each sub-project; When the progress deviation rate of a sub-project exceeds the preset warning threshold, a progress warning message is generated, which includes the sub-project name, deviation rate value and deviation cause analysis; The project progress status is judged based on the positive and negative values ​​of the progress deviation rate: When the progress deviation rate is positive, it means that the project is ahead of schedule; When the progress deviation rate is negative, it means that the project progress is lagging behind; Generate a progress deviation analysis report, including the planned project quantity, actual project quantity, deviation rate and cumulative completion percentage of each sub-project.

Images