Research On Key Technologies Of Compound Measurement System Based On Multi-sensor Information Fusion

It has a wide application in many fields to obtain the coordinate data of workpiece surface rapidly and accurately,such as positioning and inspection of workpieces,reverse engineering,and rapid prototyping manufacturing.With the rapid development of modern industry,the requirements for product function,structure,and shape are getting higher and higher,which also puts forward higher requirements for the performance of measurement system.A single measurement sensor can no longer meet the increasing requirements of modern industry for measurement accuracy,efficiency and intelligence.Therefore,the multi-sensor measurement technology has attracted more and more attention,and has become an important development trend of modern industrial measurement technology.Combine the respectively advantages of contact measurement and line structured light vision measurement,a new compound measurement system which integrates contact probe and line structured light vision sensor is proposed in this paper.The proposed system can realize high-speed,high-precision and intelligent measurement of workpieces with complex geometry in the field of mechanical manufacturing.The key technologies such as data processing and intelligent integration of measurement information are studied thoroughly,the research contents and results are as follows:A compound measurement system integrating contact probe and line structured light vision sensor is proposed to realize high-speed,high-precision and intelligent measurement of complex workpieces.The proposed measurement system first uses optical measurement information to guide contact probe to make a small amount of contact measurement,the obtained high-precision contact measurement data are then used to compensate optical measurement data through data fusion,so as to obtain a large number of high-precision optical measurement data for CAD model reconstruction.Through the deep fusion of measurement information from different digitizers,the advantages of each digitizer are brought into full play and complementary advantages are formed,thus realizing the high-speed,high-precision and intelligent measurement of the composite measurement system.A robust normal estimation method based on neighborhood segmentation and reconstruction is proposed to estimate the normal vectors of point clouds accurately and reliably,which provides a basis for the subsequent operations on point cloud,such as segmentation and feature recognition.Aiming at the problem that the existing point cloud normal estimation methods are sensitive to noise and non-uniformity of sampling and produce smooth normal vectors near the edge,a neighborhood segmentation and reconstruction strategy is proposed to construct a consistent neighborhood for each point in the point cloud that contains only points on the same surface as the current point.Then,the normal vector of the current point can be calculated accurately with the constructed consistent neighborhood.A robust point cloud segmentation and feature recognition algorithm is proposed to extract the geometric information of the measured object accurately.The proposed algorithm adopts a new region growth strategy,which takes the standard deviation of the neighborhood normal vectors of the point as the similarity measure between the points,and selects the seed point base on the similarity of the standard deviation of the neighborhood normal vectors between the current point and the initial seed point.The proposed algorithm can effectively avoid the under-segmentation of different surface features with smooth transition,and achieve the accurate segmentation of complex point cloud data.A data fusion algorithm for free-form surface measurement based on improved B-spline curve deformation technology is proposed.On the basis of traditional energy based B-spline curve deformation technology,a new expression of curve internal energy with smoothing effect is proposed,which makes the deformation algorithm keeps the overall shape of the curve while have a smooth effect on the curve,thus improves the robustness of the measurement system to noise.The contour curve obtained by optical measurement is corrected by using the proposed curve deformation technique,the corrected curve passes through all high-precision contact measuring points,while maintaining the overall shape of the original curve,and is smoother than the original curve,therefore,the deformed curve is more accurate.The proposed data fusion method improves the accuracy of optical measurement with only a few contact measurement points,can thus realize fast and accurate measurement of free-form surface.