CN-121989344-A - Short line method prefabricated segment full-automatic die adjusting method and die adjusting system based on machine vision and point cloud

CN121989344ACN 121989344 ACN121989344 ACN 121989344ACN-121989344-A

Abstract

The invention discloses a full-automatic mold adjusting method and system for a prefabricated segment by a short line method based on machine vision and point cloud. The method comprises the steps of establishing a measurement reference and completing sensor calibration, synchronously collecting images of a mould and a parent section and a three-dimensional point cloud, carrying out registration fusion, carrying out dynamic compensation on design target geometry by combining temperature, glue layer thickness and shrinkage creep, calculating deviation of a measurement surface and a compensated target surface, extracting key characteristic deviation, constructing an optimized model which takes the minimized deviation as a target and integrates actuator motion and safety clearance constraint, solving mould adjustment quantity, adopting three-stage combined position-force double closed-loop control of coarse positioning-fine matching-steady locking, driving an actuator to carry out iterative adjustment until reaching standards, and carrying out collision prediction, safety interlocking and automatic generation of traceable reports in the whole process. The invention realizes millimeter-level geometric accuracy control, obviously shortens single-segment modular time, and has the advantages of high safety and traceability of whole-flow data.

Inventors

GAO CHEN
CHEN XIAOTING
SU YINGCHAO
HUANG JUN
ZHANG LELE
ZHANG MINGBIN
ZHANG BAO
CHEN HAIFA
GENG WEIMING
ZHOU XIAOYI
LI JIA

Assignees

中国建筑第七工程局有限公司
东南大学

Dates

Publication Date: 20260508
Application Date: 20260127

Claims (10)

1. A full-automatic model adjustment method for a prefabricated segment by a short line method based on machine vision and point cloud is characterized by comprising the following steps of S1, station initialization and multi-sensor calibration, S2, multi-source data synchronous acquisition and fusion, S3, target geometry generation and process parameter dynamic compensation, S4, shape surface and characteristic deviation calculation, S5, geometric optimization solution with multiple constraints, S6, driving and iteration adjustment modules, S2, driving and iteration adjustment modules, and S2, wherein the method comprises the steps of establishing a unified measurement standard and a coordinate system, completing calibration of vision and three-dimensional point cloud acquisition equipment, synchronously acquiring images of a mould and a parent segment and three-dimensional point cloud data, carrying out fusion processing, generating fusion point cloud and pose information under the unified coordinate system, generating target geometry based on a design model, compensating the target geometry by combining the process parameters acquired in real time, comparing the fusion point cloud with the compensated target geometry, calculating shape surface deviation and key characteristic pose deviation, S5, carrying out geometric optimization solution with multiple constraints, and driving and iteration adjustment modules according to the optimal adjustment modules, and carrying out the driving and iteration adjustment modules until the optimal adjustment modules meet the requirements, and the driving and iteration adjustment modules are carried out until the optimal adjustment modules meet the requirements, and the driving modules are subjected to the driving and the driving modules.
2. The machine vision and point cloud based full-automatic modular method for pre-fabricated segments in the short-line method according to claim 1, wherein the step S1 comprises: s11, arranging a rigid reference target group on a die and a parent section; s12, defining a world coordinate system, a station coordinate system and a sensor coordinate system, and completing conversion calibration among the coordinate systems; s13, performing combined internal and external parameter calibration and time synchronization on the industrial camera and the three-dimensional scanner.
3. The method for fully automatically adjusting the model of the prefabricated segment according to the short line method based on the machine vision and the point cloud according to claim 1 or 2, wherein the step S2 comprises the following steps: s21, acquiring an initial pose from an image by identifying the reference target group; S22, registering the point clouds of all the parts to a unified coordinate system by taking the initial pose as a guide; s23, integrating vision and the position and the pose of the point cloud by adopting a filtering or optimizing algorithm, and outputting a stable position and pose.
4. The machine vision and point cloud based full-automatic modular method for pre-fabricated segments in the short-cut process according to claim 1 or 2, wherein the process parameters obtained in step S3 in combination with real-time include at least one of the following compensations: (1) Performing thermal deformation compensation based on the linear expansion coefficient of the material according to the acquired mold temperature; (2) Applying displacement offset compensation along the joint seam method according to the thickness of the designed epoxy adhesive layer; (3) And according to the concrete age, humidity and time effect, carrying out shrinkage creep prediction compensation.
5. The method for fully automatically adjusting the model of the prefabricated segment according to the short line method based on the machine vision and the point cloud according to claim 1 or 2, wherein the step S4 comprises the following steps: s41, performing shape surface segmentation and surface fitting on the fusion point cloud to obtain a measured shape surface; S42, calculating a normal distance field from the measured surface to the compensated target surface, and taking the normal distance field as a surface deviation; S43, extracting the characteristics of the shear key and the locating hole from the point cloud, and calculating the pose deviation between the shear key and the locating hole and the target characteristic.
6. The full-automatic modular method for the prefabricated segment based on the short line method of the machine vision and the point cloud according to claim 1 or 2, wherein in the step S5, the motion constraint of the actuator comprises travel, speed and acceleration limitation, the process safety constraint comprises the minimum safety clearance constraint between components calculated based on the real time point cloud, the solving adopts a sequential quadratic programming or trust domain method, and the kinematic jacobian matrix from the actuator to the die surface is updated online in iteration.
7. The full-automatic mould adjusting method for the short line method prefabricated segment based on the machine vision and the point cloud according to claim 1 or 2 is characterized in that the step S6 comprises the step S61 of driving the executing mechanism through a multi-stage closed-loop control process, wherein the process at least comprises a quick approach stage based on initial pose and a precise adjustment stage based on shape face deviation, the step S62 of performing iterative adjustment in the precise adjustment stage by adopting a control mode of integrating position feedback and force or pressure feedback, and the step S63 of re-executing the steps S2 to S5 according to feedback data after each adjustment until the closed-loop iteration is met.
8. The full-automatic modular method for the prefabricated segment based on the short line method of the machine vision and the point cloud according to claim 1 or 2, further comprising the steps of S7 safety interlocking and quality tracing: S71, collision prediction and safety interlocking are performed in real time in the whole process of die adjustment, and a rollback strategy is executed when abnormality occurs, and S72, after die adjustment is completed, a quality report containing deviation data and technological parameters is automatically generated, and whole process data is archived.
9. A modular system for the full-automatic modular method for the stub-method prefabricated segment according to any one of claims 1 to 8 is characterized by comprising an intelligent sensing module, a data fusion and geometry calculation module, an execution and control module, a safety and interlocking module, a quality tracing module and a standard archiving and archiving operation cooperation system, wherein the intelligent sensing module comprises a multi-view industrial camera and a three-dimensional laser scanner and is used for executing multi-source data acquisition and initial pose calculation in the step S2, the data fusion and geometry calculation module is used for executing calibration and initialization in the step S1, data fusion in the step S2, target geometry generation and compensation in the step S3 and deviation calculation in the step S4, the execution and control module comprises a plurality of groups of electric drive actuators, servo drivers and industrial controllers and is used for executing optimization solution in the step S5 and closed-loop modular control in the step S6, the safety and interlocking module is used for executing collision prediction, safety interlocking and abnormal rollback in the step S7, the quality tracing module is used for executing report generation and data in the step S7, and the modules are communicated with an industrial Ethernet and a real-time coordinate system and are used for conducting archiving and a unified reference work.
10. The full-automatic modular method for the prefabricated segment of the short line method based on the machine vision and the point cloud is characterized in that an actuator in the execution and control module is a multi-point electric cylinder or a six-degree-of-freedom parallel platform, the safety and interlocking module is integrated with a regional safety scanner and an emergency stop button for interlocking, a quality acceptance report generated by the quality tracing module is related to a BIM model, a control section deviation, a convergence process and a key feature comparison result are presented in a visual mode, and all process data are encrypted and filed according to a work order number.

Description

Short line method prefabricated segment full-automatic die adjusting method and die adjusting system based on machine vision and point cloud Technical Field The invention relates to the technical field, in particular to a short-line prefabricated segment mould adjusting method. Background In the construction of the prefabricated segment by the short line method, the segment to be poured needs to be in butt joint with the parent segment at the end face, and multiple geometric indexes such as end face dislocation, height difference, joint gap uniformity, torsion angle, pore channel alignment, shear key engagement, side die contour and the like need to be met at the same time. The existing flow mainly uses sparse measurement of total stations, level meters, feelers and the like, and is assisted by manual driving jacks and screw rods for repeated trial adjustment, so that the problems that measurement information is insufficient, adjustment efficiency is low, adjustment is easy, short-distance operation risk of personnel is high, unified coordinates and time references are not available in process data, and the like are solved. The Chinese patent of invention with publication No. 2024.08.30 and CN118559861A discloses a digital twinning-based method for adjusting the precision of a short line method segment beam and a medium, wherein a first end die and a second end die are arranged on a precast segment beam field, a precast segment Liang Gongduan is arranged between the first end die and the second end die, and a corresponding model is obtained through digital twinning modeling so as to precast the precast segment beam to be spliced. And the quaternions of the characteristic points of the three-dimensional opposite sides are selected on the first end die model and the second end die model for mutual comparison so as to accurately obtain the position and shape accuracy of the first end die model and the second end die model, and then the prefabricated segment Liang Moxing is respectively compared with the corresponding three-dimensional characteristic points on the first end die model and the second end die model through selecting the quaternions of the characteristic points of the three-dimensional opposite sides so as to ensure the prefabrication accuracy of the prefabricated segment beam. The invention has the following defects that 1, measurement information is sparse, shape surface control is incomplete, continuous three-dimensional geometric information of a template and a segment full shape surface cannot be obtained by means of comparison of limited discrete characteristic points, actual deviation of complex curved surfaces, joint areas and shear key meshing surfaces is difficult to comprehensively reflect, control dead zones are easy to generate in non-characteristic areas, and 2, dynamic factors of non-integrated construction are compensated in real time, namely dynamic time-varying factors in actual construction such as steel mould temperature deformation, epoxy glue layer thickness, concrete shrinkage creep and the like are not comprehensively considered in a mould adjusting process, so that theoretical target geometry is not matched with actual working conditions, and final forming precision and long-term stability are affected. Disclosure of Invention Aiming at the technical problems, the invention provides a full-automatic short-line method prefabricated segment die-adjusting method based on machine vision and point cloud, which is used for solving the problems that in the prior art, measurement information of the short-line method segment die-adjusting method is sparse, and control dead zones are easy to generate in non-characteristic areas due to incomplete surface control. In order to achieve the above purpose, the technical scheme of the invention is realized as follows: a full-automatic model adjustment method for a prefabricated segment by a short line method based on machine vision and point cloud comprises the following steps of S1, station initialization and multi-sensor calibration, S2, synchronous acquisition and fusion of multi-source data, S3, generation of target geometry and dynamic compensation of process parameters, S4, calculation of shape face and characteristic deviation, comparison of the fusion point cloud and the compensated target geometry, calculation of shape face deviation and key characteristic pose deviation, S5, geometrical optimization and solution with multiple constraints, wherein the deviation is minimized, the motion constraint and the process safety constraint of an actuator are integrated, the optimal adjustment quantity of the die is solved, S6, driving and iterative model is adjusted according to the optimal adjustment quantity, the position of the actuator is adjusted, and the die is adjusted until the new measurement deviation meets the requirement of S2, and the iterative adjustment quantity is executed until the optimal adjustment quantity meets the requi