CN-121982654-A - Bridge pier anti-collision device displacement monitoring method and system based on three-dimensional point cloud registration

CN121982654ACN 121982654 ACN121982654 ACN 121982654ACN-121982654-A

Abstract

The invention relates to the technical field of three-dimensional point cloud monitoring, in particular to a displacement monitoring method and a system of a pier anti-collision device based on three-dimensional point cloud registration, wherein the method comprises the steps of collecting three-dimensional point cloud data flow covering pier and anti-collision device areas to obtain effective slice point cloud; the method comprises the steps of calculating a rotation correction matrix and a translation correction vector of a radar relative to a pier by utilizing the absolute static characteristic of the surface of the pier, carrying out reference correction on an effective slice point cloud, dividing the corrected anti-collision device point cloud into a plurality of independent point cloud clusters by adopting an Euclidean distance clustering algorithm, connecting centroids of adjacent units in a clockwise sequence to construct a device topology framework, calculating local topology strain indexes of all units in the device topology framework, and evaluating the connection state and the morphological distortion condition of the anti-collision device. The invention can accurately distinguish the sensor shake and the device deformation, and realize accurate monitoring.

Inventors

XU MINGCAI
SONG WENTONG

Assignees

武汉力拓桥科防撞设施有限公司

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. The bridge pier anti-collision device displacement monitoring method based on three-dimensional point cloud registration is characterized by comprising the following steps of: Collecting three-dimensional point cloud data streams covering bridge piers and anti-collision device areas, and filtering environmental noise according to a preset height threshold value and a preset density threshold value to obtain effective slice point clouds; calculating a rotation correction matrix and a translation correction vector of the radar relative to the bridge pier by utilizing the absolute static characteristic of the surface of the bridge pier, and carrying out reference correction on the effective slice point cloud to eliminate the shake error of the sensor; dividing the corrected point cloud of the anti-collision device into a plurality of independent point cloud clusters by adopting an Euclidean distance clustering algorithm, extracting the geometric centroid of each point cloud cluster, connecting centroids of adjacent units in a clockwise sequence, and constructing a device topology framework; And calculating a local topological strain index of each unit in the topological skeleton of the device, and evaluating the connection state and the morphological distortion condition of the anti-collision device according to the local topological strain index to realize monitoring of the displacement of the anti-collision device of the pier.
2. The bridge pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration according to claim 1, wherein the filtering environmental noise according to a preset height threshold and density threshold comprises: Setting a height threshold value, and reserving point cloud data in a height layer where the anti-collision device and the bridge pier are located; setting a searching radius and the minimum neighbor point number in the radius, and if the neighbor point number of a certain point is less than the minimum neighbor point number, rejecting the neighbor point as an outlier.
3. The bridge pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration according to claim 1, wherein when the effective slice point cloud is subjected to reference correction, the obtained reference correction residual is calculated The reference correction residual satisfies the relationship: ; Wherein, the The number of effective bridge piers participating in calculation is counted; a correction matrix for rotation; Is the first of the point cloud subsets of the pier surface divided from the point cloud A plurality of points; is a translation correction vector; a projection point of the corrected point on the surface of the standard model; representing the euclidean norm.
4. The bridge pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration of claim 3 wherein the process of reference correction further comprises: The rotation correction matrix is obtained by using a point-to-surface iterative nearest point algorithm and iteratively solving transformation parameters which enable the distance between an actual measurement point set and a pier standard CAD model surface patch to be minimum Translation correction vector ; If the reference corrects the residual error If the radar attitude correction is smaller than the set threshold value, judging that the radar attitude correction is successful; if the reference corrects the residual error Greater than or equal to the set threshold, the current frame data is discarded.
5. The bridge pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration of claim 1, wherein the constructing device topology framework comprises: Setting a clustering tolerance, and clustering the point clouds of the anti-collision device to obtain K independent point cloud clusters; Calculating the average value of coordinates of all points in each point cloud cluster to obtain the geometric centroid of the unit; and sequencing according to the polar angle projected by each geometric centroid on a horizontal plane, and sequentially connecting the geometric centroids of adjacent units.
6. The pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration of claim 1, wherein the local topological strain index The relation is satisfied: ; In the formula, Representing the unit number; Is the first Individual cell centroid and the first The measured connection distance between the individual cell centroids; designing standard connection spacing; Is the geometric sensitivity coefficient; an included angle of the skeleton fold line at the current node; The average value of the included angles of the left and right adjacent nodes of the unit is obtained; Representing an exponential function.
7. The bridge pier collision avoidance device displacement monitoring method based on three-dimensional point cloud registration according to claim 6, wherein the estimating the connection state and the morphological distortion condition of the collision avoidance device according to the local topological strain index comprises: if the local topological strain index is Exceeding a preset warning threshold value and mainly contributing by a linear strain term in the formula, and judging that the connection fails or breaks; if the local topological strain index is Exceeding the preset warning threshold and being mainly contributed by the index term in the formula, the impact invagination or extrusion deformation is determined.
8. The method for monitoring displacement of a pier collision avoidance device based on three-dimensional point cloud registration of claim 1 further comprising collecting, at a top view angle, a three-dimensional point cloud data stream using a solid-state lidar mounted above the pier to cover the overall profile of the collision avoidance device.
9. The method for monitoring displacement of a pier anti-collision device based on three-dimensional point cloud registration according to claim 8, further comprising the steps of generating a radar chart according to the calculated local topological strain index of each unit, and visually displaying risk states of all directions of the anti-collision device.
10. The bridge pier anti-collision device displacement monitoring system based on three-dimensional point cloud registration is characterized by comprising a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions, when executed by the processor, realize the bridge pier anti-collision device displacement monitoring method based on three-dimensional point cloud registration according to any one of claims 1-9.

Description

Bridge pier anti-collision device displacement monitoring method and system based on three-dimensional point cloud registration Technical Field The invention relates to the technical field of three-dimensional point cloud monitoring. More particularly, the invention relates to a bridge pier anti-collision device displacement monitoring method and system based on three-dimensional point cloud registration. Background The floating pier anti-collision device is generally formed by connecting dozens of independent anti-collision units end to end through flexible clips and encircling the periphery of the floating pier, and has obvious non-rigid body characteristics, and the floating pier anti-collision device can generate phenomena of integral distortion, rounding, flattening or local invagination and the like under the action of water flow, waves or impact. At present, the displacement of the floating pier anti-collision device is monitored by a monitoring method based on a laser radar or a three-dimensional camera, and the two monitoring methods are used for carrying out rigid matching on the scanned integral point cloud of the floating pier anti-collision device and a standard CAD model through an integral iterative nearest point algorithm so as to judge the displacement deviation of the floating pier anti-collision device. However, when the floating pier collision avoidance device is flattened or deformed locally, the whole point cloud cannot be overlapped with the standard circular ring model, and huge calculation errors can be caused by forced rigid registration, so that the system is mistakenly used for the device to generate whole displacement, and local dangerous deformation is covered. In addition, the monitoring radar is usually installed on the side wall of the bridge or on a repair road, and is influenced by bridge vibration and wind load, and the radar self-coordinate system can slightly shake, so that the monitoring accuracy is influenced. Disclosure of Invention In order to solve the problems of large rigid registration error and monitoring misalignment caused by sensor shaking, the invention provides a bridge pier anti-collision device displacement monitoring method and system based on three-dimensional point cloud registration, which can accurately distinguish sensor shaking and device deformation and realize accurate monitoring. In a first aspect, the invention provides a bridge pier anti-collision device displacement monitoring method based on three-dimensional point cloud registration, which comprises the steps of collecting three-dimensional point cloud data streams covering bridge piers and an anti-collision device region, filtering environmental noise according to a preset height threshold value and a preset density threshold value to obtain effective slice point clouds, calculating a rotation correction matrix and a translation correction vector of a radar relative to the bridge piers by utilizing absolute static characteristics of the bridge pier surface, carrying out reference correction on the effective slice point clouds, eliminating sensor shake errors, dividing the corrected anti-collision device point clouds into a plurality of independent point cloud clusters by adopting an Euclidean distance clustering algorithm, extracting geometric centroids of each point cloud cluster, connecting centroids of adjacent units in a clockwise sequence to construct a device topology skeleton, calculating local topological strain indexes of all units in the device topology skeleton, and evaluating the connection state and the morphological distortion condition of the anti-collision device according to the local topological strain indexes to realize monitoring on the displacement of the bridge pier anti-collision device. By adopting the technical scheme, the three-dimensional point cloud data flow covering the bridge pier and the anti-collision device area is collected, noise is filtered, the absolute static characteristic of the bridge pier surface is utilized to carry out reference correction on the radar attitude, a topological skeleton reflecting the device form is constructed, the local topological strain index is calculated, the self-shaking error of the sensor can be accurately stripped in a high-dynamic water surface environment, the connection state and the form distortion of the flexible chain structure are effectively monitored, and therefore the accuracy and the anti-interference capability of monitoring data are remarkably improved. Preferably, the filtering environmental noise according to the preset height threshold and density threshold comprises setting a height threshold, reserving point cloud data in a height layer where an anti-collision device and a pier are located, setting a searching radius and the minimum neighbor point number in the radius, and if the neighbor point number of a certain point is less than the minimum neighbor point number, eliminating the point as an outlier