Research On Key Technologies For Automated Measurement Of Automotive Complex Parts

It is an inevitable tendency to manufacture cars with more customization and intellectualization in transformation and upgrading of automotive industry.For those auto companies,the tradeoff between requirements from customers and fabrication quality is a great challenge.For irregular parts measurement,complex features have to be measured automatically with much local surface detail,high precision and high efficiency.Existing methods couldn’t satisfy all measurement requirements.As a result,it is an urgent task to find an automatic measurement method to measure complex parts efficiently and precisely.Aiming to measure complex parts in cars manufacture automatically,we propose a high precision method combining terminal scanning measurement and global accuracy control location in this work.Point clouds are obtained in car scale with high precision and high density.Industrial robots and guidance based on machine vision realize automatic measurement and logistics together.Measured features are evaluated correctly by referring to car body design principle and three-coordinate measuring evaluation.We study several important technologies in this work such as terminal measurement,global accuracy control with dual geometric constraint and automatic measurement based on machine vision.A system experiment shows feasibility of our proposed methods.The main contents show as follows:1.The development direction of automotive industry and the importance of key car-body parts measurement are summarized.Features of car-body parts measurement are analyzed in detail.We introduce our implementation and methods by studying contradiction between measurement precision and efficiency.2.We study high-precision structured light projection by introducing phase unwrapping,three-dimensional reconstruction,internal parameters calibration and distortion correction in cameras and projectors.External observation method of terminal measurement and its calibration are designed and verified experimentally.3.We build a global accuracy control model based on dual geometric constraint to uniform measurement data from terminal measurements.Precision unification from terminal measurements to global coordinate system is accomplished.Then we study rapid calibration method for external parameters of multi cameras based on robot kinematics and hand-eye relation.4.Automatic measurement and logistics based on the vision localization are proposed after analyzing the importance of automation in measurement process.We build a vision localization model and propose initial value estimation,error analysis and precision optimization methods.Reliability and precision of those methods are proved by experiments.5.Aiming to evaluate complex features in car body rapidly,we propose point cloud fusion and index modeling methods to improve segmentation precision of data.Different evaluation algorithms are designed for different features and proved by experiments.