Research On Registration Algorithms In 3D Reconstruction

Three dimensional laser scanning technology is a new technology,and application of digital city,cultural relics protection and reverse engineering.But at present,the technology is still in the initial stage of the point cloud data processing,there are many areas to be improved.Especially in the face of huge point cloud data,in order to obtain high precision and high quality results,it is needed to seek a kind of point cloud data processing method which is highly efficient,stable and highly automatic.So it is very necessary to research the point cloud data processing of 3D laser scanning technology and develop the system and the efficient point cloud data processing method.The object of this paper is the three-dimensional model reconstruction of the three-dimensional laser technology,the focus of the three-dimensional reconstruction process of the point cloud data registration issues related research and experiments,the main work is as follows:1 Summarizes the 3D laser model processes,and its steps are discussed and studied in detail,and experiments were performed using the existing point cloud,that the method can establish accurate point cloud model complete and efficient.2 Research on the registration algorithm,focuses on the analysis of the traditional iterative closest point algorithm(ICP),aiming at the shortcomings of traditional ICP algorithm to the high initial position requirements and low searching efficiency,this paper introduces the initial registration and accurate registration registration strategy.At the time,in the initial registration,the invariant feature is used to carry on the registration of the point to meet the initial registration requirement.In the accurate registration algorithm using distance limit to calculate the internal bit set to reduce the workload and the search algorithm,the traditional ICP algorithm into k-d tree in order to improve the matching point of search efficiency,and the use of point cloud matching relationship between given point to restrictive conditions to improve the registration accuracy of point cloud data.3 based on the improved registration algorithm,in the Microsoft Visual Studio2013 environment using C++,using PCL1.8.0 version of point cloud library implementation algorithm.And using the point cloud data to compare with the previous algorithms,it is proved that the improved algorithm can effectively improve the efficiency and accuracy of point cloud registration.On the basis of this,the GeomagicStudio2012 is used to reconstruct the 3D model according to the reconstruction process listed in this paper.