Research On Key Technique Of Viewpoint Generation And Path Planning For Automated Surface Structured-light 3D Measurement

One kind of non-contact optical 3D measurement is Surface structured-light 3D measurement which has the advantages of high precision,high speed and high resolution.So it has been widely used in industrial detection,reverse engineering,biomedical and other fields,and has achieved good application.Nevertheless,the traditional equipments for surface structured-light 3D measuring require manual operation during measuring process,and is difficult to meet the needs of high-speed,high-volume detection.It is a trend for future digital factories to develop high-performance sensors that combine surface structured-light equipments with industrial robots to automate surface structured-light 3D measurement.The difficulty of automating surface structured-light 3D measurement focus on the planning of measuring viewpoints and path.At this stage,measuring viewpoints and path are edited by manual teaching,so it is quite time-consuming and hardly to obtain best viewpoints and path which can improve measuring efficiency.In view of above problems,this paper thoroughly researches the techniques of measuring viewpoint and path planning.On the basics of existing viewpoint planning technique,this paper introduces one Generate & Test method to generate viewpoints automatically.Firstly,according to surface information of complex NURBS model and measurement requirements,the model surface partition is established with the Measurement Space as constraint,and several candidate viewpoints are then generated which will be filter by measuring quality constraint involving optical visibility,imaging quality and surface characteristics.At last the highest measuring quality viewpoint is selected as best viewpoint.Supported by the viewpoint generation technique,this paper further studies the problem of planning path by generated viewpoints,and introduces one method to plan measuring path.The method creatively abstracts the set of generated viewpoints as a weighted undirected graph with the robot motion fees as the weight,which converts the path planning problem into the TSP problem searching smallest weight Hamilton circuit in the graph.Based on the Greedy Theory,the TSP problem is decomposed into three subproblems including to search the shortest path in the set of specified viewpoints,to search the shortest from the start point to the set of specified viewpoints,to search the shortest path from the set of specified viewpoints to the end point.And these three shortest path compose the best path.Thanks to above two methods,one planning system for automated surface structuredlight 3D measurement is developed on RobotStudio platform.According to reference model,this system is able to automatically generate measuring viewpoints and plan measuring path.Furthermore,the system is practically applied to the planning for automate measurement on steering knuckles beltline.Experiments have shown that the planning speed of the planning system is fast,and the measuring efficiency of planned viewpoints and path is high.It can replace manual teaching in the practical process.