A Random Order Point Set Matching Algorithm for Industrial Measurement Comparative Analysis

In industrial measuring application, common point pairs must be manually selected for coordinate system transformation in comparative analysis on measured point set and designed point set. Aiming at this problem, a comparative analysis algorithm are proposed on the basis of random order point set matching and principle of least square, in order to realize automatic matching and comparative analysis of precision in this kind of application and acquire optimum precision control. At first, two incompletely corresponding unordered point sets in different coordinate systems are taken as processed objects, and point pair matching of measurement point set and design point set is automated through analysis on relative topological relationship of space, so as to realize coordinate system transformation computation. Then all point pairs are applied and filtered according to elementary matching deviation, converting parameters are recalculated to improve precision of overall coordinate transformation, and measurement points with errors are acquired. Experiments and analysis prove validity of the algorithm and its application conditions are discussed.