Research On Improved Algorithm Of Point Cloud Registration Based On ICP Algorithm

Along with the development of 3D laser scanning technology,more and more applied in the construction of national production,and this technology has some advantages compared with the traditional measurement,for example:Big data?Data acquisition speed?High accuracy?Strong initiative?All time?Real time?Information transmission,processing,easy to express,etc.More and more attention from the majority of the people.Along with the rapid development of remote scanning technology in the last few years,it has expanded its application field.In the field of engineering survey,digital city,mine surveying,urban planning,ancient protection,virtual reality and so on has a broad application prospect.Aiming at the problem of low accuracy and robustness of the registration of the mid point cloud registration in 3D modeling.According to the principle of the classical ICP(Iterative Closest Point)algorithm,two aspects of the coarse registration and fine registration have improved.In the coarse registration of PCA principal component analysis method,RANSAC algorithm,geometric features of the coarse registration algorithm in the algorithm for the high degree of ovcrlap,the cost of the algorithm is so large,the point of view of the higher number of cloud geometrical properties.A registration algorithm based on edge feature point cloud is proposed.The speed of fine registration part for the classical ICP algorithm is slow,there exists the wrong point,a number of robustness problems.Using the improved iterative function,kd-tree(K-dimensional Tree)acceleration iteration and CPC(Closest Point Criterion)three constraint method,the algorithm is improved.The algorithm is analyzed and compared,and the results are obtained.In the coarse registration,there is no obvious point cloud,which can deal with the low degree of overlap and geometrical features.In practical engineering application and get better results.