CN-121685872-B - Building pipeline node auxiliary positioning method applying three-dimensional digital management

CN121685872BCN 121685872 BCN121685872 BCN 121685872BCN-121685872-B

Abstract

The invention discloses a building pipeline node auxiliary positioning method applying three-dimensional digital management, which relates to the technical field of building construction and comprises the following steps of firstly collecting multi-source data of a multi-equipment terminal, fusing the multi-source data through self-adaptive federal filtering, outputting three-dimensional coordinates and component parameters of nodes, secondly, obtaining BIM attribute diagrams of building pipelines in advance, constructing actual measurement attribute diagrams based on the three-dimensional coordinates and the component parameters of the nodes, screening deviation nodes by matching spatial positions and semantic attributes of the BIM attribute diagrams and the actual measurement attribute diagrams through a graph convolution network, correcting the deviation nodes by reinforcement learning, outputting a corrected calibration model, thirdly, automatically highlighting corresponding pipeline components in BIM based on the three-dimensional coordinates and the calibration model of the nodes, superposing and displaying design parameters and historical maintenance records, and realizing intelligent and iterative correction of the deviation nodes by utilizing a depth Q network according to an automatic deviation type and grade decision correction strategy, so that manual intervention is greatly reduced.

Inventors

FENG XINGLONG
ZHANG ZONGWEI
Zu Jinwei
ZHANG YAFEI

Assignees

西安星讯智能通信科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260210

Claims (5)

1. The building pipeline node auxiliary positioning method applying three-dimensional digital management is characterized by comprising the following steps of: The method comprises the steps of firstly, collecting multi-source data of a multi-equipment terminal, fusing the multi-source data through self-adaptive federal filtering, and outputting three-dimensional coordinates and component parameters of nodes; The BIM attribute map of the building pipeline is obtained in advance, an actual measurement attribute map is constructed based on three-dimensional coordinates of nodes and component parameters, spatial positions and semantic attributes of the BIM attribute map and the actual measurement attribute map are matched through a map convolution network, so that deviation nodes are screened out, the deviation nodes are corrected through reinforcement learning, and a corrected calibration model is output; The process for constructing the actually measured attribute graph comprises the following steps: acquiring three-dimensional coordinates and component parameters of nodes, associating the three-dimensional coordinates with actual measurement nodes acquired on site, restoring pipeline geometric information and a topology network based on association relations, acquiring semantic tags of the actual measurement nodes, and storing and constructing an actual measurement attribute graph in a graph structure by combining the component parameters; the process of screening out the deviation node is as follows: aligning node attribute fields of BIM attribute graphs and actually measured attribute graphs, eliminating numerical difference through min-max standardization, constructing topological relation by using adjacent matrixes, distributing node unique IDs, adopting network structures of an input layer, a 2-layer graph convolution layer and an output layer to extract high-dimensional feature vectors fusing space, semantics and topological information for the two types of graphs respectively, calculating the similarity of the space position and the similarity of the semantic attribute between the nodes, weighting according to space weight coefficients and semantic weight coefficients to obtain comprehensive matching degree, and judging matched nodes: When the comprehensive matching degree is greater than or equal to a matching degree threshold value, judging as a matching node; When the comprehensive matching degree is smaller than the matching degree threshold value, judging as unmatched nodes, dismantling specific differences of space positions and semantic attributes, judging as deviation nodes, dividing deviation types and grades, quantifying as deviation grades, and screening the deviation nodes; The process of resolving the specific differences of the spatial position and the semantic attribute is as follows: When the single-dimensional difference is disassembled, for nodes with the comprehensive matching degree not reaching the standard, respectively checking the spatial position similarity and the semantic attribute similarity of the nodes; when the deviation level is quantized, calculating the Euclidean distance of the three-dimensional coordinates of the actually measured node and the corresponding BIM node, namely the space linear distance, and dividing the deviation level according to the distance; Summarizing all nodes determined to be space deviation, semantic deviation and compound deviation in the deviation node screening output process, and eliminating invalid difference nodes caused by data acquisition errors; The process of outputting the corrected calibration model is as follows: Acquiring a BIM attribute map, an actual measurement attribute map and a deviation node list as a state space, and correcting BIM parameters, data recovery and marking abnormality as an action space and deviation elimination rate as a reward function; When space deviation exists, correcting BIM coordinates by slight deviation, checking data first for moderate deviation, correcting and marking reasons, marking construction abnormality by serious deviation, generating correction suggestion, when semantic deviation exists, updating BIM semantic attribute by slight deviation, executing updating after serious deviation triggers rechecking, when compound deviation exists, making a decision according to high-level deviation priority, synchronously correcting BIM coordinates and semantics by slight deviation, calling a graph convolution network to be re-matched, judging that the deviation elimination rate is more than or equal to 90%, otherwise, adjusting strategy to be re-corrected, integrating BIM and measured data after effective correction, and forming a calibration model; And thirdly, automatically highlighting corresponding pipeline components in the BIM based on the three-dimensional coordinates of the nodes and the calibration model, and displaying design parameters and historical maintenance records in a superposition mode.
2. The aided positioning method of building pipeline node with three-dimensional digital management according to claim 1, wherein the process of collecting multi-source data of the multi-device terminal is: and deploying terminals of integrated Beidou differential positioning, MEMS inertial navigation and electromagnetic signal acquisition modules in areas corresponding to the building pipeline nodes, respectively acquiring three-dimensional coordinates, positions, postures and electromagnetic fingerprints of the pipeline nodes, and respectively preprocessing to obtain multi-source data.
3. The aided positioning method of building pipeline node with three-dimensional digital management of claim 2, wherein the process of outputting the three-dimensional coordinates and the component parameters of the node is: Calculating a positioning coordinate and a covariance matrix by using the three-dimensional coordinates of the pipeline nodes as state quantity and combining the satellite quantity and the signal-to-noise ratio to output accuracy confidence coefficient; the position and the gesture of the pipeline node are used as state quantity, the recursive coordinate and the covariance matrix are updated through Kalman filtering, and the confidence coefficient is dynamically adjusted according to the drift characteristic; and distributing weights based on the confidence coefficient, fusing the state estimation values to obtain final node three-dimensional coordinates, and matching component parameters by combining electromagnetic fingerprints and BIM pre-stored information.
4. The method for assisted positioning of building pipeline nodes using three-dimensional digital management according to claim 1, wherein the process of highlighting the corresponding pipeline components in the BIM is: The method comprises the steps of carrying a positioning mobile terminal with a preassembled auxiliary positioning system into an area, acquiring three-dimensional coordinates of personnel through a multi-source fusion positioning module configured by the positioning mobile terminal, synchronously calling a calibration model, matching the three-dimensional coordinates of the personnel with the coordinates of pipeline nodes of the model to screen nodes, associating corresponding pipeline components based on a topological relation of the model to form positioning and component mapping, positioning and amplifying a view of the calibration model to the association area at a mobile terminal interface, filling and contour thickening highlight target components with high contrast colors, superposing virtual models of the target components to a real scene, and carrying out luminous highlighting rendering.
5. The aided positioning method of building pipeline node with three-dimensional digital management of claim 1, wherein the process of superposing and displaying design parameters and historical maintenance records is as follows: on the basis of highlighting the component, extracting semantic attributes of the component from the calibration model, displaying the semantic attributes beside the component in a floating information frame, synchronously connecting an operation and maintenance database, and calling a history record through a unique ID of the component to supplement the information frame.

Description

Building pipeline node auxiliary positioning method applying three-dimensional digital management Technical Field The invention relates to the technical field of building construction, in particular to a building pipeline node auxiliary positioning method applying three-dimensional digital management. Background In the field of building engineering, a pipeline system is one of core infrastructures for realizing building functions, such as water supply and drainage, heating ventilation and electric pipelines, but the existing pipeline node management has various pain points, namely on one hand, pipelines are buried in walls, ceilings or underground, space information deviation easily occurs depending on a positioning mode of a two-dimensional drawing, so that the pipeline is difficult to search in a construction stage and a maintenance stage, on the other hand, design parameters and maintenance records of the pipeline nodes are stored in paper documents or scattered electronic forms, the information is inconvenient to call in on-site operation, and the problems of low operation and maintenance efficiency, hysteresis of fault treatment and the like are easily caused. While BIM is widely applied to pipeline design, the problem of digital and physical disconnection between the existing BIM model and the actual working condition on site is remarkable, the BIM model is mostly static data in the design stage, the actual pipeline position and the component state after construction are difficult to match in real time, and meanwhile, the single positioning technology has insufficient precision in complex environments such as indoor and underground, and the accurate positioning of pipeline nodes cannot be realized. Disclosure of Invention In order to achieve the above purpose, the invention is realized by the following technical scheme: The method comprises the steps of firstly, collecting multi-source data of a multi-equipment terminal, fusing the multi-source data through self-adaptive federal filtering, and outputting three-dimensional coordinates and component parameters of nodes; The BIM attribute map of the building pipeline is obtained in advance, an actual measurement attribute map is constructed based on three-dimensional coordinates of nodes and component parameters, spatial positions and semantic attributes of the BIM attribute map and the actual measurement attribute map are matched through a map convolution network, so that deviation nodes are screened out, the deviation nodes are corrected through reinforcement learning, and a corrected calibration model is output; And thirdly, automatically highlighting corresponding pipeline components in the BIM based on the three-dimensional coordinates of the nodes and the calibration model, and displaying design parameters and historical maintenance records in a superposition mode. Further, the process of collecting multi-source data of the multi-device terminal is as follows: and deploying terminals of integrated Beidou differential positioning, MEMS inertial navigation and electromagnetic signal acquisition modules in areas corresponding to the building pipeline nodes, respectively acquiring three-dimensional coordinates, positions, postures and electromagnetic fingerprints of the pipeline nodes, and respectively preprocessing to obtain multi-source data. Further, the process of outputting the three-dimensional coordinates of the node and the component parameters is as follows: Calculating a positioning coordinate and a covariance matrix by using the three-dimensional coordinates of the pipeline nodes as state quantity and combining the satellite quantity and the signal-to-noise ratio to output accuracy confidence coefficient; the position and the gesture of the pipeline node are used as state quantity, the recursive coordinate and the covariance matrix are updated through Kalman filtering, and the confidence coefficient is dynamically adjusted according to the drift characteristic; and distributing weights based on the confidence coefficient, fusing the state estimation values to obtain final node three-dimensional coordinates, and matching component parameters by combining electromagnetic fingerprints and BIM pre-stored information. Further, the process of constructing the measured attribute map is as follows: and acquiring three-dimensional coordinates and component parameters of the nodes, associating the three-dimensional coordinates with the actually measured nodes acquired on site, restoring pipeline geometric information and a topology network based on association relations, acquiring semantic tags of the actually measured nodes, and storing and constructing an actually measured attribute graph in a graph structure by combining the component parameters. Further, the process of screening out the deviation node is as follows: aligning node attribute fields of BIM attribute graphs and actually measured attribute graphs, eliminating numerical difference throu