Research Of Data Matching And Registration On The Surface Of Micro Pyramid Arrays On Turning Tool

Grinding the micro-structured arrays on the surface of turning tool can improve theefficiency of scraps discharge and heat release, which will improve the performance ofturning tool. But the shape of three-dimensional structure at micrometer scale is hard to bemeasured, and the machining accuracy can not be controlled. It hinders to develop thecharacters of performance further. Therefore, the study will evaluate the machining accuracyof micro pyramid arrays with433μm height on the surface of turning tool which is processedby micro grinding technology and measured by white light interferometer.The measurement range of white light interferometer is only300μm, so that the pointclouds of micro pyramid have to split into upper and lower parts after measurement. Since itis not consistent between the processing coordinate system and the measurement system,Iterative Closest Point algorithm (ICP) is used to make two parts matching and registrationand analyze the matching error. First, there are a certain noises in measured point cloud. Thede-noising method of neighbors at different points and threshold is used to analyze theperformance of de-noising. Then using these de-noising data matches the theory model byinitiative matching and ICP algorithm matching, and determines the best de-noising method.Finally, the direct registration can get the whole micro pyramid after the data matching. Touse the method of the partial coincidence point cloud matches the theory model again by ICPalgorithm. It makes the two parts closer and gets the better result of registration.The research results show that the average error and standard error of matching willconverge to a stable value, and the average error can be called micro-structured machiningform error. When there is no initiative matching, the matching times will increase and reducethe efficiency of matching, but it will not affect the final matching result. After ICP algorithmmatching, the minimum average error of lower micro pyramid is14.33μm by de-noisingmethod of2-point neighbors at2.0μm threshold and the minimum average error of uppermicro pyramid is12.87μm by de-noising method of12-point neighbors at1.0μm threshold.The ICP matching by partial coincidence point cloud can improve the result of registration.Both the average error and the standard error are important for matching analysis and thestandard error will converge to a small and stable value at registration. The minimum standarderror of registration can be achieved through matching the point cloud of coincidence and60%extension. The micro-structured machined form error for the whole micro pyramid is14.42μm.