A digital supervision method and system for power construction projects

Abstract

This invention relates to the field of digital supervision technology, specifically a digital supervision method and system for power construction projects, comprising the following steps: collecting surface three-dimensional coordinate points and clustering them to calculate component point cloud clusters; applying a random sampling consensus algorithm to extract the measured center axis; comparing the design axis network coordinates to calculate the axis network deviation value; constructing an axis network topology diagram; traversing and comparing deviation values; analyzing and marking primary deviation sources; and recursively calculating and determining the deviation propagation path. In this invention, by collecting three-dimensional coordinate points and clustering them, the measured center axis of the component is calculated to restore its geometric shape. The spatial translation vector and rotation matrix are then compared with the design axis network to quantify objective deviation values, eliminating subjective human intervention. Simultaneously, the component connections are used to construct an axis network topology diagram. Combined with the recursive calculation of deviation values, the source of deviation is effectively analyzed and marked, and the deviation propagation path is determined, achieving closed-loop management from deviation discovery to source diagnosis.

Description

Technical Field

[0001] This invention relates to the field of digital supervision technology, and in particular to a digital supervision method and system for power construction projects. Background Technology

[0002] Digital monitoring technology is a comprehensive field that utilizes information technology to monitor, collect, analyze, and manage specific objects or processes in real time. Its core aspects include automatically acquiring on-site status information through various sensors, cameras, drones, and other devices; remotely transmitting data using the Internet of Things and mobile communication networks; and summarizing, storing, and visualizing information through a data platform. It systematically integrates information collection terminals, network communication links, and backend management platforms.

[0003] Traditional digital monitoring systems for power construction projects refer to technical solutions that monitor personnel safety, equipment status, and project progress at construction sites. They typically employ the following methods: for personnel safety, location information is obtained through GPS chips embedded in safety helmets, and key areas are manually inspected using fixed video surveillance cameras; for equipment status, operators manually fill out equipment operation logs via handheld terminal apps, or small tools are registered for entry and exit using RFID tags; for project progress, on-site management personnel take photos and upload them to project management software, which are then manually compared with construction drawings to determine progress.

[0004] Existing technologies have significant limitations in supervising construction quality. They mainly rely on manual observation and manual data reporting, such as comparing photos taken by on-site personnel with drawings to judge progress, or manually filling in equipment status through an app. This method is highly dependent on the experience and sense of responsibility of the operators, making it difficult to guarantee the objectivity and timeliness of the data. Video monitoring is also limited to manual inspection of key areas and cannot quantitatively analyze the precise spatial position and geometric shape of construction components. When construction deviations occur, traditional methods can only detect superficial inconsistencies, such as visually observing a deviation in the position of a component, but cannot accurately calculate the specific value of the deviation, let alone trace the initial installation error that caused the deviation, or determine the transmission effect of this error on subsequent processes, making it difficult to eliminate potential quality problems at the source. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a digital supervision method and system for power construction projects.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a digital supervision method for power construction projects, comprising the following steps:

[0007] S1: Collect a set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set a neighborhood determination threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood determination threshold, and calculate the component point cloud cluster;

[0008] S2: For the point cloud cluster of the component, set the number of iterations and the tolerance value, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, and calculate the measured central axis;

[0009] S3: Call the measured center axis and the designed grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, and define the spatial translation vector and rotation matrix as the grid deviation value;

[0010] S4: Construct a network topology diagram based on the component connection relationship, set an initial deviation threshold, call the network topology diagram and the network deviation value, traverse the nodes and compare the network deviation value, analyze and mark the nodes that are greater than the initial deviation threshold and greater than the network deviation value of the adjacent nodes as first-level deviation sources, call the first-level deviation sources to recursively calculate and determine the actual transmission path of the deviation.

[0011] As a further embodiment of the present invention, the component point cloud cluster includes a cluster identifier, cluster center point coordinates, and number of points within the cluster; the measured central axis specifically includes an axis direction vector and coordinates of a reference point on the axis; the axis network deviation value includes X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis; the axis network topology diagram includes a topology node set and a topology edge set; the primary deviation source specifically refers to the deviation source node identifier; and the actual propagation path of the deviation is specifically an ordered node sequence.

[0012] As a further aspect of the present invention, step S1 specifically comprises:

[0013] Collect a set of three-dimensional coordinate points on the surface to obtain the relationship between the design grid coordinates and the component connections;

[0014] A neighborhood determination threshold is set, and a K-dimensional tree structure is used to spatially index the set of three-dimensional coordinate points on the surface.

[0015] Based on the neighborhood determination threshold, a density reachability search is performed to divide the set of three-dimensional coordinate points on the surface into multiple point cloud clusters. Target point cloud clusters are selected and identified according to the component connection relationship. Cluster identifiers, cluster center coordinates, and number of points within the cluster are calculated to generate component point cloud clusters.

[0016] As a further aspect of the present invention, step S2 specifically comprises:

[0017] Call the component point cloud cluster, and set the number of iterations and tolerance values;

[0018] Within the number of iterations, the smallest point set is randomly selected from the component point cloud cluster to construct the initial geometric primitive;

[0019] Calculate the distance from all points in the component point cloud cluster to the initial geometric primitive, count the points whose distance is less than the tolerance value as interior points, and calculate the number of interior points;

[0020] Compare the current number of interior points with the historical maximum number of interior point supports, and retain the initial geometric primitive with the maximum number of interior point supports as a candidate geometric primitive;

[0021] Extract the spatial equation of the central axis of the candidate geometric primitive, calculate the direction vector of the axis and the coordinates of the reference point on the axis, and generate the measured central axis.

[0022] As a further aspect of the present invention, step S3 specifically comprises:

[0023] Call the measured center axis and the designed grid coordinates;

[0024] Extract the direction vector of the measured center axis, the coordinates of the reference point on the axis, and the direction vector and reference point coordinates of the designed axis grid.

[0025] The optimal rigid body transformation relationship between the measured central axis and the designed grid coordinates is calculated using the singular value decomposition algorithm.

[0026] Separate the translation and rotation components from the optimal rigid body transformation relationship, calculate the X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis, combine the translation components into a spatial translation vector, and combine the rotation components into a rotation matrix to generate the axis grid deviation value.

[0027] As a further aspect of the present invention, step S4 specifically comprises:

[0028] By invoking the component connection relationships, a axial network topology diagram including the topology node set and the topology edge set is constructed;

[0029] Set an initial deviation threshold, call the axis network topology diagram and the axis network deviation value, traverse all nodes in the topology node set, compare the axis network deviation value of each node with the initial deviation threshold, and compare its axis network deviation value with the axis network deviation value of the adjacent nodes;

[0030] Analyze and mark nodes whose axis grid deviation value is greater than the initial deviation threshold and greater than the axis grid deviation value of all adjacent nodes as first-level deviation sources, and obtain the deviation source node identifier;

[0031] Starting from the first-level deviation source, a depth-first search is performed along the topological edge set of the axis network topology diagram to recursively calculate the propagation effect of the deviation between adjacent nodes;

[0032] The nodes traversed by the depth-first search are recorded to form the ordered node sequence, and the actual propagation path of the deviation is generated.

[0033] As a further aspect of the present invention, the step of calculating the distance from all points in the component point cloud cluster to the initial geometric primitive specifically comprises:

[0034] For any point in the component point cloud cluster Its coordinates are ;

[0035] The spatial equation of the central axis of the initial geometric primitive is obtained, defined as follows: ;

[0036] Calculation points The spatial equation of the central axis Euclidean distance ;

[0037] The The calculation formula is:

[0038] ;

[0039] in, The spatial equation representing the central axis The axis direction vector, The spatial equation representing the central axis The coordinates of the reference point on the axis are as follows. Representative point The position vector, For the index of the point, The Euclidean norm of a vector. Represents the cross product of vectors;

[0040] Comparison Compared with the tolerance value, if If the value is less than the tolerance value, then the determination point is... Let be the interior point.

[0041] As a further aspect of the present invention, the step of calculating the optimal rigid body transformation relationship between the measured central axis and the designed grid coordinates using the singular value decomposition algorithm specifically comprises:

[0042] Extract from the measured center axis and the designed grid coordinates respectively Each corresponding feature point forms a set of measured points. With design point set ;

[0043] Calculate the measured point set center of mass With the design point set center of mass ;

[0044] Construct the covariance matrix:

[0045] ;

[0046] in, For the measured point set The point in the middle, For the design point set The corresponding point in the middle, For the index of the feature point, The total number of feature points. Represents the transpose of a matrix;

[0047] For the covariance matrix Perform singular value decomposition ;

[0048] Calculate the rotation matrix Calculate the spatial translation vector The rotation matrix With the spatial translation vector Together they constitute the optimal rigid body transformation relationship.

[0049] As a further aspect of the present invention, the step of performing a depth-first search along the topological edge set of the network topology diagram and recursively calculating the propagation effect of the deviation between adjacent nodes specifically comprises:

[0050] Create a set of visited nodes and a transmission path stack. Push the identifier of the deviation source node of the first-level deviation source onto the transmission path stack and add it to the set of visited nodes.

[0051] Obtain the node at the top of the transmission path stack as the current node, and find the downstream adjacent node in the axis network topology diagram that is connected to the current node and is not in the set of visited nodes;

[0052] If the downstream adjacent node is captured, and the axis network deviation value of the current node is greater than the axis network deviation value of the downstream adjacent node, and the axis network deviation value of the downstream adjacent node is greater than the initial deviation threshold;

[0053] Then, the downstream adjacent node is pushed onto the propagation path stack, added to the visited node set, and the edges in the topological edge set from the current node to the downstream adjacent node are recorded in the ordered node sequence;

[0054] If no downstream adjacent node that meets the conditions is captured, the current node is popped from the propagation path stack, completing one backtracking.

[0055] A digital monitoring system for power construction projects, wherein the digital monitoring system for power construction projects is used to implement the aforementioned digital monitoring method for power construction projects, the system comprising:

[0056] The point cloud clustering module is used to collect a set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set a neighborhood determination threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood determination threshold, calculate the component point cloud clusters, and transmit them to the axis extraction module.

[0057] The axis extraction module is used to set the number of iterations and tolerance values ​​for the component point cloud cluster, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, calculate the measured central axis, and pass it to the spatial deviation calculation module.

[0058] The spatial deviation calculation module is used to call the measured center axis and the design grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, define the spatial translation vector and the rotation matrix as the grid deviation value, and pass it to the deviation transmission analysis module;

[0059] The deviation propagation analysis module is used to construct a network topology diagram through the connection relationship of the components, set an initial deviation threshold, call the network topology diagram and the network deviation value, traverse the nodes and compare the network deviation value, analyze and mark the nodes that are greater than the initial deviation threshold and greater than the network deviation value of the adjacent nodes as primary deviation sources, call the primary deviation sources to recursively calculate and determine the actual propagation path of the deviation.

[0060] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0061] In this invention, by collecting a set of three-dimensional coordinate points on the surface and clustering them with the coordinates of the design grid and the connection relationship of the components, the measured central axis of the component point cloud cluster is calculated, thereby achieving accurate geometric restoration of the construction entity. Then, through spatial geometric comparison, the spatial translation vector and rotation matrix of the measured axis and the design grid are automatically calculated to quantify the grid deviation value. This method gets rid of the subjectivity of traditional manual visual inspection or photo comparison, provides objective deviation data, and uses the connection relationship of the components to construct the grid topology diagram. Combined with the deviation value, the nodes are traversed for recursive calculation, which can effectively analyze and mark the source of deviation, determine the actual transmission path of deviation, and realize closed-loop management from the discovery of deviation to the diagnosis of the source. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the overall process of the digital supervision method of the present invention.

[0063] Figure 2 This is a flowchart illustrating the point cloud cluster generation process of the present invention.

[0064] Figure 3 This is a flowchart of the actual measurement and calculation of the center axis of the present invention;

[0065] Figure 4 This is a flowchart of the calculation of the axis grid deviation value of the present invention;

[0066] Figure 5 This is a flowchart for determining the actual transmission path of deviation in this invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of this invention clearer, the software-based technical solution is described in detail below with reference to system architecture diagrams and embodiments. It should be understood that the specific embodiments described herein are only for explaining the technical solutions of this invention and do not constitute a limitation on the scope of protection.

[0068] In the description of this invention, the system architecture relationships or data processing flows indicated by terms such as "layer," "module," "interface," "data flow," "client," and "server" are all defined based on the architecture diagram or flowchart corresponding to the embodiments. This way of describing is only used to clearly illustrate the logical relationships between the elements in the technical solution, and not to limit the physical deployment form. The term "multiple" includes two or more technical units, including but not limited to multiple data nodes, processing threads, service instances, or functional components and other scalable elements. The specific number is determined according to the actual business scenario and needs to be specifically specified.

[0069] Please see Figure 1 and Figure 2 This invention provides a technical solution: a digital supervision method for power construction projects, comprising the following steps:

[0070] S1: Collect the set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set the neighborhood judgment threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood judgment threshold, and calculate the component point cloud cluster;

[0071] A component point cloud cluster includes a cluster identifier, cluster center coordinates, and the number of points within the cluster.

[0072] Collect a set of three-dimensional coordinate points on the surface to obtain the relationship between the design grid coordinates and the component connections;

[0073] A neighborhood determination threshold is set, and a K-dimensional tree structure is used to spatially index the set of three-dimensional coordinate points on the surface.

[0074] Density reachability search is performed based on the neighborhood determination threshold. The set of three-dimensional coordinate points on the surface is divided into multiple point cloud clusters. The target point cloud clusters are selected and identified according to the component connection relationship. The cluster identifier, the coordinates of the cluster center point and the number of points in the cluster are calculated to generate the component point cloud cluster.

[0075] A 3D laser scanner was used to scan key load-bearing structures in the power construction project (e.g., foundation base N1, steel component N2, and steel component N3) to obtain raw point cloud data covering their surfaces. This dataset contains 5,000,000 3D coordinate points. ,in The coordinate unit is meters. For example, point Coordinates are

[0076] Extract the design grid coordinates of the target component "N2 steel component" from the project's Building Information Model (BIM) or CAD design drawings. Define this as a baseline, with the coordinates of the baseline point as follows: The direction vector is At the same time, the connection relationship of the components is extracted. The upstream (connecting) component of steel component N2 is "foundation base N1", and the downstream (connecting) component is "beam N4".

[0077] A neighborhood determination threshold is set, and a K-dimensional tree structure is used to spatially index the set of three-dimensional coordinate points on the surface.

[0078] Set neighborhood determination threshold The threshold is set based on the minimum spacing between components and the average density of the scanned point cloud.

[0079] To ensure that point cloud clusters of different components can be effectively separated while maintaining the integrity of the point cloud of a single component, a neighborhood determination threshold verification experiment was conducted. Five different parameter combinations were used in the experiment, and the results are shown in Table 1.

[0080] Table 1: Results of the neighborhood determination threshold verification experiment

[0081] ;

[0082] As shown in Table 1, when the threshold is set to 0.05 meters or 0.10 meters, a single component is incorrectly segmented into multiple point cloud clusters (oversegmentation). When the threshold is set to 0.50 meters, the point clouds of adjacent components N2 and N4 are incorrectly merged into the same cluster (undersegmentation). The results of experiments 3 and 5 indicate that setting a neighborhood determination threshold is necessary. The meter can achieve accurate segmentation under different point cloud densities.

[0083] As shown in Table 1, when the threshold is set to 0.05 meters or 0.10 meters, a single component is incorrectly segmented into multiple point cloud clusters (oversegmentation). When the threshold is set to 0.50 meters, the point clouds of adjacent components N2 and N4 are incorrectly merged into the same cluster (undersegmentation). The results of experiments 3 and 5 indicate that setting a neighborhood determination threshold is necessary. The meter can achieve accurate segmentation under different point cloud densities.

[0084] Density reachability search is performed based on the neighborhood determination threshold. The set of three-dimensional coordinate points on the surface is divided into multiple point cloud clusters. The target point cloud clusters are selected and identified according to the component connection relationship. The cluster identifier, the coordinates of the cluster center point and the number of points in the cluster are calculated to generate the component point cloud cluster.

[0085] Set the minimum number of points for the core object in density search. .

[0086] Iterate through all points and select one unvisited point.

[0087] Using a K-dimensional tree, search of - Neighborhood ( All points within a radius of meters.

[0088] Query the K-dimensional tree and return the set of neighborhood points. Count the number of points in its neighborhood. .

[0089] judge ,determination Create a new cluster with the core node as the core.

[0090] Will All 35 points were added to the pending processing queue. .

[0091] from Take a point from Query of - Neighborhood ( A set of points within a radius of meters The number of points is .

[0092] judge ,determination It is also the core point. Its neighborhood... Unvisited points that do not belong to any cluster are added to the queue. .

[0093] If from Points extracted from If the number of neighboring points is less than 10, then mark it as a boundary point and classify it as a boundary point.

[0094] Repeat this process until the queue is full. Empty, cluster Build complete.

[0095] Select the next unvisited point and repeat the search until all points have been visited.

[0096] This process divides the 5,000,000 points into 15 point cloud clusters.

[0097] Calculate the coordinates of the center point of each cluster (the average coordinates of all points within the cluster).

[0098] Center point:

[0099] Center point:

[0100] Center point:

[0101] …

[0102] Center point:

[0103] Will center point Design grid reference point of "N2 steel component" Perform spatial location comparison, Euclidean distance between the two Meters (this is the average center point on the component, not a reference point). Meanwhile... center point It is close to the designed location of "N1 Foundation Base"; center point It is close to the design position of "N4 crossbeam".

[0104] Based on the connection relationship between "Steel Component N2" and "Foundation Base N1" and "Beam N4", select the locations spatially related to the designed positions of N1 and N4 (i.e., and Adjacent

[0105] logo For the target point cloud cluster.

[0106] calculate The number of points within the cluster, .

[0107] calculate The coordinates of the cluster center point,

[0108] calculate The cluster identifier is "Cluster_N2_BeamA".

[0109] Generate component point cloud clusters (cluster identifier: "Cluster_N2_BeamA", cluster center coordinates: , the number of points in the cluster: 85320).

[0110] Please see Figure 1 and Figure 3 S2: For the component point cloud cluster, set the number of iterations and tolerance value, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, and calculate the measured central axis.

[0111] The measured center axis specifically consists of the axis direction vector and the coordinates of the reference point on the axis;

[0112] Call the component point cloud cluster, and set the number of iterations and tolerance values;

[0113] Within the number of iterations, the minimum point set is randomly selected from the component point cloud cluster to construct the initial geometric primitive;

[0114] Calculate the distance from all points in the component point cloud cluster to the initial geometric primitive, count the points whose distance is less than the tolerance value as interior points, and calculate the number of interior points;

[0115] Compare the current number of inliers with the historical maximum number of inlier supports, and retain the initial geometric primitive with the maximum number of inlier supports as candidate geometric primitives;

[0116] Extract the spatial equation of the central axis of the candidate geometric primitives, calculate the axis direction vector and the coordinates of the reference point on the axis, and generate the measured central axis.

[0117] For any point in the component point cloud cluster Its coordinates are ;

[0118] The spatial equation of the central axis of the initial geometric primitive is obtained, defined as follows: ;

[0119] Calculation points equation of space to the central axis Euclidean distance ;

[0120] The calculation formula is:

[0121] ;

[0122] in, Representing the spatial equation of the central axis The axis direction vector, Representing the spatial equation of the central axis The coordinates of the reference point on the axis. Representative point The position vector, For the index of the point, The Euclidean norm of a vector. Represents the cross product of vectors;

[0123] Compare With tolerance value, if If it is less than the tolerance value, then the decision point is... Let be an interior point.

[0124] Call the component point cloud cluster "Cluster_N2_BeamA" (containing 85,320 points) generated by S1.

[0125] Set the number of iterations The basis for setting this number of times is to ensure a high level of confidence (e.g., Select at least once the minimum set of points with no outliers. Assume the proportion of outliers is... Then the interior point ratio .calculate To ensure the stability of the algorithm's convergence, the number of iterations is set. Second-rate.

[0126] Set tolerance value This value is set with reference to the nominal accuracy of the scanning device (3D laser scanner) (e.g., The surface flatness of the construction components (steel components) was measured in mm. To determine the optimal tolerance value, tests were conducted on the point cloud of a calibrated column with known dimensions.

[0127] 1. Settings Meters: The RANSAC algorithm identifies 32% of the inliers, which is too few and the axis fitting is unstable.

[0128] 2. Settings Meters: The RANSAC algorithm identifies 85% of the interior points, the axis fit is stable, and the deviation from the calibration value is minimal.

[0129] 3. Settings The RANSAC algorithm identified 96% of the interior points, but it included too many noise points from component edges, causing the axis fitting results to deviate from the calibration value. Based on the above experiments, a tolerance value was set. rice.

[0130] The random sampling consensus algorithm is used to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, and calculate the measured central axis.

[0131] Initialize the maximum number of inliers in history .

[0132] Execute the first iteration ( ):

[0133] Two points are randomly selected from 85,320 points to form the minimum point set: and

[0134] Constructing the initial geometric primitives (central axis) ):

[0135] coordinates of reference point on axis

[0136] Axis direction vector

[0137] Iterate through 85,320 points and calculate the value of each point. arrive distance

[0138] statistics The number of points in a meter (number of interior points). .

[0139] Compare and . ,renew ,reserve As a current candidate geometric primitive.

[0140] Perform the second iteration ( ):

[0141] Random selection (This point is an outlier) and

[0142] Build : ,

[0143] Traverse all points and calculate arrive The distance.

[0144] statistics The number of interior points in meters. .

[0145] Compare and . ,throw away

[0146] …

[0147] Perform the 240th iteration ( ):

[0148] Random selection and

[0149] Build .

[0150] Count the number of in-points. .

[0151] Compare and . ,renew ,reserve As candidate geometric primitives.

[0152] …

[0153] Perform the 500th iteration ( ):

[0154] The number of iterations reached Stop iterating.

[0155] With the largest number of interior support points ( )of It was selected as the final candidate geometric primitive.

[0156] (Optional refinement steps: use) By refitting the axis using principal component analysis (PCA) or least squares method from all (52,150) interior points, a more accurate axis equation can be obtained.

[0157] Extracting candidate geometric primitives The spatial equation parameters of the central axis (after refinement):

[0158] Axis direction vector (The vector has been normalized)

[0159] coordinates of reference point on axis .

[0160] Generate the measured center axis (axis direction vector:

[0161] Coordinates of the reference point on the axis: ).

[0162] against The formula, explained in detail below:

[0163] formula .

[0164] Parameter description: The first representative component in the point cloud cluster Points To the initial geometric primitive (central axis) The Euclidean distance of ). yes The axis direction vector. yes The coordinates of the reference point on the axis. It is a point The position vector, It is the index of the point. This represents the cross product operation of vectors in three-dimensional space. It represents the Euclidean norm of a vector (i.e., the magnitude of the vector).

[0165] Operational logic: A reference point for the axis is calculated. Point to the test point The vector. A new vector is calculated that is perpendicular to the axis. sum vector According to the geometric meaning of the vector cross product, the norm of the new vector... Numerically equal to that of a vector and The area of ​​the parallelogram formed by the adjacent sides. It is the norm of the axial direction vector, i.e., the length of the base of the parallelogram. It is calculated by dividing the area of ​​the parallelogram by the length of its base. That is, we obtain the height of the parallelogram, which is the point. To the axis vertical distance .

[0166] Parameter acquisition and calculation examples:

[0167] In the first iteration, the parameters were obtained as follows:

[0168] .

[0169] .

[0170] Select any point from the component point cloud cluster. ( ) to perform calculations, .

[0171] Tolerance value rice.

[0172] Substitute into the calculation:

[0173] 1. .

[0174] 2. .

[0175] 3. .

[0176] 4. .

[0177] 5. Meters. Result comparison: Calculated... Rice. (The rest of the text appears to be unrelated and possibly machine-generated.) With tolerance value (0.020 meters) are compared. The result is true. This indicates that the point... To the initial geometric primitive The distance is less than the set tolerance value. Therefore, during the interior point statistics process of the first iteration, the point... It was determined to be an interior point. The count increases by 1. The advantage of the formula is that it allows for... , and The cross product and norm operations between points provide a computationally efficient and geometrically meaningful way to calculate the "point-to-line distance" in three-dimensional space, independent of a specific coordinate system. This allows the RANSAC algorithm to quickly and accurately evaluate how well each randomly generated initial geometric primitive (central axis) fits all points in a point cloud cluster, thus robustly extracting the set of interior points from point cloud data containing a large amount of noise and outliers.

[0178] Please see Figure 1 and Figure 4S3: Call the measured center axis and the design grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, and define the spatial translation vector and rotation matrix as the grid deviation value;

[0179] The grid deviation values ​​include X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis;

[0180] Use the measured center axis and the designed grid coordinates;

[0181] Extract the axis direction vector of the measured center axis, the direction vector of the reference point coordinates on the axis and the coordinates of the design axis grid;

[0182] The singular value decomposition algorithm is used to calculate the optimal rigid body transformation relationship between the measured central axis and the design grid coordinates.

[0183] Separate the translation and rotation components from the optimal rigid body transformation relationship, calculate the X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis, combine the translation components into a spatial translation vector, combine the rotation components into a rotation matrix, and generate the axis grid deviation value;

[0184] Extracting coordinates from the measured center axis and the designed grid lines respectively Each corresponding feature point forms a set of measured points. With design point set ;

[0185] Calculate the measured point set center of mass With design point set center of mass ;

[0186] Construct the covariance matrix:

[0187] ;

[0188] in, For the measured point set The point in the middle, Design point set The corresponding point in the middle, For the index of the feature point, The total number of feature points. Represents the transpose of a matrix;

[0189] For covariance matrix Perform singular value decomposition ;

[0190] Calculate the rotation matrix Calculate the spatial translation vector Rotation matrix With spatial translation vector Together they constitute the optimal rigid body transformation relationship.

[0191] Call the measured center axis generated by S2 (for "N2 steel component"):

[0192] ;

[0193] (Unit: meters)

[0194] Call the design grid coordinates extracted by S1 (for "N2 steel component"):

[0195] ;

[0196] (Unit: meters)

[0197] Perform spatial geometric comparison, calculate spatial translation vectors and rotation matrices, and define spatial translation vectors and rotation matrices as axis grid deviation values; extract coordinates from measured center axis and designed axis grid. Each corresponding feature point forms a set of measured points. With design point set .

[0198] measured point set Extraction:

[0199] ;

[0200] ;

[0201] ;

[0202] Design Points Extraction:

[0203] ;

[0204] ;

[0205] ;

[0206] Calculate the measured point set center of mass :

[0207] ;

[0208] ;

[0209] ;

[0210] ;

[0211] Calculate the design point set center of mass :

[0212] ;

[0213] ;

[0214] ;

[0215] ;

[0216] Constructing the covariance matrix .

[0217] Calculate the decentralized point:

[0218] ;

[0219] ;

[0220] ;

[0221] ;

[0222] ;

[0223] ;

[0224] calculate :

[0225] ;

[0226] ;

[0227] ;

[0228] ;

[0229] ;

[0230] For covariance matrix Perform singular value decomposition And calculate the rotation matrix. This step is performed using a numerical calculation library and aims to obtain the design point set. Rotate to the measured point set The optimal rotation. In this embodiment, the calculated rotation matrix... (Approximately a rotation of 0.286 degrees around the Y-axis) is:

[0231] ;

[0232] Calculate the spatial translation vector .

[0233] ;

[0234] ;

[0235] ;

[0236] ;

[0237] ;

[0238] ;

[0239] Separate translation and rotation components:

[0240] Spatial translation vector (Unit: meters)

[0241] X-axis translation: -48 mm (rounded up for use in subsequent tables).

[0242] Y-axis translation: -54 mm.

[0243] Z-axis translation: -148 mm (rounded up for use in subsequent tables).

[0244] Rotation matrix correspond:

[0245] Rotation angle around the X-axis: 0 degrees.

[0246] Rotation angle around the Y-axis: 0.286 degrees.

[0247] Rotation angle around the Z-axis: 0 degrees.

[0248] Combined translation components and rotational components Generate grid deviation values ​​(X-axis translation: -48mm, Y-axis translation: -54mm, Z-axis translation: -148mm, rotation angle around X-axis: 0°, rotation angle around Y-axis: 0.286°, rotation angle around Z-axis: 0°).

[0249] against The formula, explained in detail below:

[0250] formula .

[0251] Parameter description: yes The covariance matrix. It is the set of measured points The first in One feature point, It is a design point set Zhongyu The corresponding number 1 feature point. It is the set of measured points The center of mass, that is . It is a design point set The center of mass, that is . It is the index of the feature point. It is the total number of feature points (in this example) ). This represents the transpose operation of a matrix.

[0252] Operational logic: and Calculate the "decentralized" position vectors of the measured point and the design point relative to their respective centroids. (column vector). Calculate the outer product of these two decentralized vectors, i.e., one Vector multiplied by one Vector, to obtain a matrix. All Each corresponding point pair was calculated The matrix is ​​summed element by element. This summation process constructs a... A matrix that encodes two "decentralized" point sets ( and Rotational correlation between ( ).

[0253] Parameter acquisition and calculation examples:

[0254] .

[0255] .

[0256] .

[0257] , .

[0258] , .

[0259] , .

[0260] Substitute into the calculation:

[0261] ;

[0262] ;

[0263] ;

[0264] ;

[0265] Comparison of results:

[0266] The calculation result is the covariance matrix. .

[0267] The result It is A matrix. It is not a directly comparable scalar in itself. This matrix is ​​the input to Singular Value Decomposition (SVD). The purpose of SVD is to extract values ​​from the matrix. middle( Decomposed into and matrix, The matrix calculation is used to obtain the rotational components in the optimal rigid body transformation relationship. ( This is a necessary intermediate step.

[0268] The advantage of the formula lies in its ability to construct the covariance matrix. It will find the optimal rotation between two sets of points. The problem is transformed into a linear algebraic problem that can be stably solved by singular value decomposition (SVD). The SVD decomposition of the matrix provides a closed solution ( This solution guarantees the optimal orthogonal rotation, thus avoiding the local optimum or iteration non-convergence problems that may occur in nonlinear optimization methods, and ensuring the accuracy and robustness of rigid body transformation calculation.

[0269] Please see Figure 1 and Figure 5 S4: Construct a grid topology diagram based on component connection relationships, set an initial deviation threshold, call the grid topology diagram and grid deviation value, traverse nodes to compare grid deviation value, analyze and mark nodes that are greater than the initial deviation threshold and greater than the grid deviation value of adjacent nodes as first-level deviation sources, call the first-level deviation source to recursively calculate, and determine the actual propagation path of the deviation.

[0270] The axis network topology diagram includes a set of topological nodes and a set of topological edges;

[0271] The primary deviation source specifically refers to the deviation source node identifier;

[0272] The actual propagation path of the deviation is specifically an ordered sequence of nodes;

[0273] By invoking the component connection relationships, a network topology diagram including a set of topological nodes and a set of topological edges is constructed;

[0274] Set an initial deviation threshold, call the axis network topology diagram and axis network deviation value, traverse all nodes in the topology node set, compare the axis network deviation value of each node with the initial deviation threshold, and compare its axis network deviation value with the adjacent nodes.

[0275] Analyze and mark nodes whose grid deviation value is greater than the initial deviation threshold and greater than the grid deviation value of all adjacent nodes as first-level deviation sources, and obtain the deviation source node identifier;

[0276] Starting from the primary deviation source, a depth-first search is performed along the topological edge set of the axis network topology diagram to recursively calculate the propagation effect of the deviation between adjacent nodes;

[0277] Record the nodes visited during the depth-first search to form an ordered node sequence and generate the actual propagation path of the deviation;

[0278] Create a set of visited nodes and a propagation path stack. Push the deviation source node identifier of the first-level deviation source onto the propagation path stack and add it to the set of visited nodes.

[0279] Get the node at the top of the transmission path stack as the current node, and find the downstream adjacent node in the axis network topology graph that is connected to the current node and is not in the set of visited nodes.

[0280] If the downstream adjacent node is captured, and the current node's axis network deviation value is greater than the downstream adjacent node's axis network deviation value, and the downstream adjacent node's axis network deviation value is greater than the initial deviation threshold;

[0281] Then push the downstream adjacent nodes onto the propagation path stack, add them to the set of visited nodes, and record the edges in the topological edge set from the current node to the downstream adjacent nodes into the ordered node sequence.

[0282] If no downstream adjacent node that meets the conditions is captured, the current node is popped from the propagation path stack, and a backtracking is completed.

[0283] The component connection relationship extracted from S1 is invoked, which describes the physical connection sequence between foundations, steel components and beams in various power construction projects;

[0284] Table 2: Component Connection Relationship Table

[0285] ;

[0286] As shown in Table 2, the axial network topology diagram is constructed based on the component connection relationships. .

[0287] Topology node set .

[0288] Topological edge set , where (N1, N2) means that N1 is the upstream node of N2.

[0289] Set initial deviation threshold The threshold is set with reference to the allowable deviations for steel structure grid installation in the "Standard for Quality Inspection and Evaluation of Power Engineering Construction" (e.g., DL / T5168-2018). The standard stipulates that the translational deviation of the main load-bearing components should not exceed 15mm, and the verticality deviation should not exceed [a certain value]. And no more than 25mm.

[0290] To conduct preliminary screening, an initial deviation threshold for comprehensive translational deviation is set. Millimeters.

[0291] The algorithm calls the axis network topology diagram and axis network deviation values, traverses nodes to compare axis network deviation values, analyzes and marks nodes that are greater than the initial deviation threshold and greater than the axis network deviation values ​​of their adjacent nodes as primary deviation sources, and recursively calculates the actual propagation path of the deviation by calling the primary deviation source algorithm; for the topology node set... For each node (N1 to N6), repeat steps S1 to S3 to calculate the respective axis deviation value.

[0292] For ease of comparison, the "maximum translation deviation" is used. The value of the axis deviation of the node is the maximum value of the absolute value of the translation of the X, Y, and Z axes calculated in S3;

[0293] Table 3: Calculation Results of Grid Deviation Values ​​for Each Node

[0294] ;

[0295] As shown in Table 3, the deviation values ​​of node N2-BeamA (-48, -54, -148) are related to the spatial translation vector calculated in S3. Meters (rounded to the nearest whole number, corresponding to X, Y, and Z axis translations of -48mm, -54mm, and -148mm respectively), with its maximum translation deviation. mm. The deviation values ​​of other nodes (N1, N3-N6) were measured and calculated using the same method.

[0296] Traversing the topology node set :

[0297] 1. Check node N1 (Base): mm, Check the adjacent (downstream) nodes of N1: N2 ( ) and N3 ( ), (real), (True), node N1 satisfies "greater than the initial deviation threshold" and "greater than the axis deviation value of all adjacent nodes", and is marked as a first-level deviation source.

[0298] 2. Check node N2 (BeamA):

[0299] mm, Check the adjacent nodes of N2: N1( ) and N4 ( ), (False), N2 does not meet the condition.

[0300] 3. Check node N3 (BeamB):

[0301] mm, Check the adjacent nodes of N3: N1( ) and N5 ( ),

[0302] (False), N3 does not meet the condition, ... (N4, N5, and N6 also do not meet the condition because they all have at least one neighboring node with a deviation value greater than itself). Traversal ends, obtain the deviation source node identifier. Starting from the first-level deviation source N1, perform a depth-first search (DFS) along the topological edge set of the axis network topology graph: Initialization: set of visited nodes Transmission path stack ordered node sequence Press N1 in , N1 added , Iteration 1: Current node Find the downstream nodes of N1: N2, N3, and process N2 (unvisited): , , ,Compare ( ), if true, compare ( If the condition is true, then N2 is pushed in. , N2 added Record the edge (N1, N2) to Iteration 2: Current node Find the downstream node of N2: N4, and process N4 (unvisited): , ,Compare (True), compared (True), condition met, N4 will be pushed in. , N4 added Record the edge (N2, N4) to Node before iteration Find the downstream node of N4: N6, and process N6 (not visited): , ,Compare (True), compared (True), condition met, N6 will be pushed in. , N6 added Record edge (N4, N6) to Iteration 4: Current node Find downstream nodes of N6: None. Backtrack, from... N6 pops up. Iteration 5: Current node Find the downstream node (N6) of N4. N6 is already... In the middle, there are no unvisited downstream nodes. Backtracking, N4 is popped. Iteration 6: Current node Find the downstream node (N4) of N2. N4 is already... In the middle, backtrack, and N2 is popped. Iteration 7: Current node Find the downstream nodes (N2, N3) of N1. N2 has been visited. Process N3 (not visited): , ,Compare (True), compared (True), condition met, N3 will be pushed in. , N3 added Record the edge (N1, N3) to Iteration 8: Current node Find the downstream node of N3: N5, and process N5 (not visited): , ,Compare (True), compared (True), condition met, push N5 in. , ,

[0303] N5 added Record edge (N3, N5) to Iteration 9: Current node Find the downstream node of N5: N6, process N6 (already visited): N6 is already... In the middle, backtrack, N5 pops up. Iteration 10: Current node No unvisited downstream nodes. Backtrack, N3 is popped. Iteration 11: Current node No unvisited downstream nodes, backtrack, pop N1. The stack is empty; the search has ended. The ordered node sequence recorded in the middle is: [(N1,N2),(N2,N4),(N4,N6),(N1,N3),(N3,N5)], and the actual propagation path of the generated deviation is: [(N1,N2),(N2,N4),(N4,N6),(N1,N3),(N3,N5)].

[0304] A digital monitoring system for power construction projects, used to execute the aforementioned digital monitoring method for power construction projects, includes:

[0305] The point cloud clustering module is used to collect a set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set a neighborhood determination threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood determination threshold, calculate the component point cloud clusters, and transmit them to the axis extraction module.

[0306] The axis extraction module is used to set the number of iterations and tolerance values ​​for the point cloud clusters of components, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, calculate the measured central axis, and pass it to the spatial deviation calculation module.

[0307] The spatial deviation calculation module is used to call the measured center axis and the design grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, define the spatial translation vector and rotation matrix as the grid deviation value, and pass it to the deviation transmission analysis module;

[0308] The deviation propagation analysis module is used to construct a grid topology diagram based on component connection relationships, set an initial deviation threshold, call the grid topology diagram and the grid deviation value, traverse the nodes to compare the grid deviation value, analyze and mark nodes that are greater than the initial deviation threshold and greater than the grid deviation value of adjacent nodes as primary deviation sources, call the primary deviation sources to recursively calculate, and determine the actual propagation path of the deviation.

[0309] The above embodiments illustrate preferred embodiments of the present invention. Any equivalent adjustments to the technical solution based on software engineering methods are within the scope of protection, including but not limited to: implementing algorithm logic using different programming languages, refactoring functional modules into services, adjusting data interaction protocols, and optimizing resource scheduling strategies. Any implementation scheme derived from reasonable modifications to the data processing flow, service call chain, or system architecture layer without departing from the core technology of the present invention should be considered within the scope of protection defined by the claims of the present invention.

Claims

1. A digital supervision method for power construction projects, characterized in that, Includes the following steps: S1: Collect a set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set a neighborhood determination threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood determination threshold, and calculate the component point cloud cluster; S2: For the point cloud cluster of the component, set the number of iterations and the tolerance value, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, and calculate the measured central axis; S3: Call the measured center axis and the designed grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, and define the spatial translation vector and rotation matrix as the grid deviation value; S4: Construct an axis network topology diagram based on the component connection relationship, set an initial deviation threshold, call the axis network topology diagram and the axis network deviation value, traverse the nodes and compare the axis network deviation value, analyze and mark the nodes that are greater than the initial deviation threshold and greater than the axis network deviation value of the adjacent nodes as first-level deviation sources, call the first-level deviation sources to recursively calculate and determine the actual transmission path of the deviation; The specific steps of S4 are as follows: By invoking the component connection relationships, a axial network topology diagram including a set of topological nodes and a set of topological edges is constructed; Set an initial deviation threshold, call the axis network topology diagram and the axis network deviation value, traverse all nodes in the topology node set, compare the axis network deviation value of each node with the initial deviation threshold, and compare its axis network deviation value with the axis network deviation value of the adjacent nodes; Analyze and mark nodes whose axis grid deviation value is greater than the initial deviation threshold and greater than the axis grid deviation value of all adjacent nodes as primary deviation sources, and obtain the node identifier of the primary deviation source; Starting from the first-level deviation source, a depth-first search is performed along the topological edge set of the axis network topology diagram to recursively calculate the propagation effect of the deviation between adjacent nodes; Record the nodes visited during the depth-first search to form an ordered node sequence and generate the actual propagation path of the deviation; The specific steps of performing a depth-first search along the topological edge set of the network topology graph and recursively calculating the propagation effect of the deviation between adjacent nodes are as follows: Create a set of visited nodes and a transmission path stack. Push the identifier of the deviation source node of the first-level deviation source onto the transmission path stack and add it to the set of visited nodes. Obtain the node at the top of the transmission path stack as the current node, and find the downstream adjacent node in the axis network topology diagram that is connected to the current node and is not in the set of visited nodes; If the downstream adjacent node is captured, and the axis network deviation value of the current node is greater than the axis network deviation value of the downstream adjacent node, and the axis network deviation value of the downstream adjacent node is greater than the initial deviation threshold; Then, the downstream adjacent node is pushed onto the propagation path stack, added to the visited node set, and the edges in the topological edge set from the current node to the downstream adjacent node are recorded in the ordered node sequence; If no downstream adjacent node that meets the conditions is captured, the current node is popped from the propagation path stack, completing one backtracking.

2. The digital supervision method for power construction projects according to claim 1, characterized in that, The component point cloud cluster includes a cluster identifier, cluster center point coordinates, and number of points within the cluster. The measured central axis specifically includes the axis direction vector and the coordinates of the reference point on the axis. The axis network deviation values ​​include X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis. The axis network topology diagram includes a topology node set and a topology edge set. The primary deviation source specifically refers to the deviation source node identifier. The actual propagation path of the deviation is specifically an ordered node sequence.

3. The digital supervision method for power construction projects according to claim 1, characterized in that, The specific steps of S1 are as follows: Collect a set of three-dimensional coordinate points on the surface to obtain the relationship between the design grid coordinates and the component connections; A neighborhood determination threshold is set, and a K-dimensional tree structure is used to spatially index the set of three-dimensional coordinate points on the surface. Based on the neighborhood determination threshold, a density reachability search is performed to divide the set of three-dimensional coordinate points on the surface into multiple point cloud clusters. Target point cloud clusters are selected and identified according to the component connection relationship. Cluster identifiers, cluster center coordinates, and number of points within the cluster are calculated to generate component point cloud clusters.

4. The digital supervision method for power construction projects according to claim 1, characterized in that, The specific steps of S2 are as follows: Call the component point cloud cluster, and set the number of iterations and tolerance values; Within the number of iterations, the smallest point set is randomly selected from the component point cloud cluster to construct the initial geometric primitive; Calculate the distance from all points in the component point cloud cluster to the initial geometric primitive, count the points whose distance is less than the tolerance value as interior points, and calculate the number of interior points; Compare the current number of interior points with the historical maximum number of interior point supports, and retain the initial geometric primitive with the maximum number of interior point supports as a candidate geometric primitive; Extract the spatial equation of the central axis of the candidate geometric primitive, calculate the direction vector of the central axis and the coordinates of the reference point on the axis, and generate the measured central axis.

5. The digital supervision method for power construction projects according to claim 2, characterized in that, The specific steps of S3 are as follows: Call the measured center axis and the designed grid coordinates; Extract the direction vector of the measured center axis, the coordinates of the reference point on the axis, and the direction vector and reference point coordinates of the designed axis grid. The optimal rigid body transformation relationship between the measured central axis and the designed grid coordinates is calculated using the singular value decomposition algorithm. Separate the translation and rotation components from the optimal rigid body transformation relationship, calculate the X-axis translation, Y-axis translation, Z-axis translation, rotation angle around the X-axis, rotation angle around the Y-axis, and rotation angle around the Z-axis, combine the translation components into a spatial translation vector, and combine the rotation components into a rotation matrix to generate the axis grid deviation value.

6. The digital supervision method for power construction projects according to claim 4, characterized in that, The specific steps for calculating the distance from all points in the component point cloud cluster to the initial geometric primitive are as follows: For any point in the component point cloud cluster Its coordinates are ; The spatial equation of the central axis of the initial geometric primitive is obtained, defined as follows: ; Calculation points The spatial equation of the central axis Euclidean distance ; The The calculation formula is: ； in, The spatial equation representing the central axis The axis direction vector, The spatial equation representing the central axis The coordinates of the reference point on the axis are as follows. Representative point The position vector, For the index of the point, The Euclidean norm of a vector. Represents the cross product of vectors; Comparison Compared with the tolerance value, if If the value is less than the tolerance value, then the determination point is... Let be the interior point.

7. The digital supervision method for power construction projects according to claim 5, characterized in that, The steps for calculating the optimal rigid body transformation relationship between the measured central axis and the designed grid coordinates using the singular value decomposition algorithm are as follows: Extract from the measured center axis and the designed grid coordinates respectively Each corresponding feature point forms a set of measured points. With design point set ; Calculate the measured point set center of mass With the design point set center of mass ; Construct the covariance matrix: ； in, For the measured point set The point in the middle, For the design point set The corresponding point in the middle, For the index of the feature point, The total number of feature points. Represents the transpose of a matrix; For the covariance matrix Perform singular value decomposition ; Calculate the rotation matrix Calculate the spatial translation vector The rotation matrix With the spatial translation vector Together they constitute the optimal rigid body transformation relationship.

8. A digital monitoring system for power construction projects, characterized in that, The system is used to implement the digital supervision method for power construction projects as described in any one of claims 1-7, and the system includes: The point cloud clustering module is used to collect a set of three-dimensional coordinate points on the surface, extract the design grid coordinates and the connection relationship between the components, set a neighborhood determination threshold, cluster the set of three-dimensional coordinate points on the surface based on the neighborhood determination threshold, calculate the component point cloud clusters, and transmit them to the axis extraction module. The axis extraction module is used to set the number of iterations and tolerance values ​​for the component point cloud cluster, apply the random sampling consensus algorithm to search for candidate geometric primitives with the maximum number of interior point support, extract the spatial equation of the central axis of the candidate geometric primitives, calculate the measured central axis, and pass it to the spatial deviation calculation module. The spatial deviation calculation module is used to call the measured center axis and the design grid coordinates, perform spatial geometric comparison, calculate the spatial translation vector and rotation matrix, define the spatial translation vector and the rotation matrix as the grid deviation value, and pass it to the deviation transmission analysis module; The deviation propagation analysis module is used to construct a network topology diagram through the connection relationship of the components, set an initial deviation threshold, call the network topology diagram and the network deviation value, traverse the nodes and compare the network deviation value, analyze and mark the nodes that are greater than the initial deviation threshold and greater than the network deviation value of the adjacent nodes as primary deviation sources, call the primary deviation sources to recursively calculate and determine the actual propagation path of the deviation.

Images