3D Laser Radar Measurement System Development And Research On Its Measurement Uncertainty

Three-dimensional(3D)laser scanning technique has been widely used for acquiring the profile of the object by scanning the surface from a far distance with a high speed.The data acquired is called the point cloud from which the dimensional information about the object can be obtained through data processing and analyzing.Characterized as high efficiency and non-contact measurement,3D laser scanning technique is especially favored in the industrial measurement field,where the measurement site is too terrible or the workpiece to be measured is too large for the traditional contact-measurement method to be applied to.However,during the scanning process,the point cloud data acquired is inevitably subject to the comprehensive influence of multiple error factors,resulting in the measurement error in the final result.Therefore,it is necessary to provide a reliable uncertainty assessment for the 3D laser scanning system.The current researches are mostly limited to the error source analysis of the scanning system or the uncertainty of the single measurement point,while few researches can offer an effective measurement uncertainty evaluation method for the geometric features in the point cloud.In most cases,the most commonly-used approach is still the traditional one,such as the repeated measurement method,which is inefficient and time-consuming.Therefore,this paper is aimed to propose a novel method for uncertainty evaluation of the geometric features in the point cloud based on a Time-ofFlight(TOF)3D laser radar measurement system.The main work of this paper includes:1)Establishment of the single-point uncertainty model and the spatial error distribution map for the 3D laser radar measurement system.The main error sources of the measurement system were analyzed and the positional error distribution of the single measurement point was studied through error propagation.Then based on the error ellipsoid theory the single-point uncertainty model was developed.Meanwhile,the spatial error distribution map for the measurement system was constructed.2)Measurement uncertainty evaluation study for different geometric features based on the error ellipsoid theory.Based on the single-point uncertainty model,the point cloud error distribution map for multiple different geometric features,such as the plane,the cylinder,etc.were constructed.Then based on the RANSAC fitting principle,the “critical point” theory was proposed prompting the establishment of measurement uncertainty evaluation models for multiple geometric features,specifically speaking,including the point-plane distance,the plane-plane distance and the cylindrical diameter,etc.Also,in terms of every procedure of the measurement from data acquisition to data processing,corresponding strategies were explored to improve the measurement accuracy of this system.Experiments were designed and conducted for the verification of these uncertainty models and the error-diminishing strategies.3)Software development of the 3D laser radar measuring system.The software was developed by LabVIEW integrating with C++.LabVIEW is used to develop the motor controlling model and the radar communicating model for the point cloud data acquisition,whereas the C++ together with PCL was implemented for the data processing model.The data process mainly includes point cloud filtering,segmentation,feature detection and dimension measurement,etc.