Three Dimensional Point Cloud The Neighborhood Search And Filtering Algorithm Reserch

3D reconstruction become increasingly important in the information time.It is the bridge between computer and three-dimensional world.In many domains such as medical,aerospace, computer vision has received widespread attention.Neighborhood search algorithm and filtering algorithm can effectively simplify and optimize the initial point cloud model.The preparing of the 3D reconstruction is the most basic and crucial step in the technology.This paper proposes a adaptive neighborhood search window algorithm and a filtering algorithm that based on Gaussian filtering algorithm.In this paper, three-dimensional reconstruction of neighborhood search algorithms and filtering algorithm is studied.The main contents are as follows:First of all,to research the original point cloud data.Due to the different density of point cloud data,and the data from equipment on the junction between the foreground and the background exist great fluctuation, and a lot of depth information missing.Proposed algorithm, an adaptive neighborhood search window for neighborhood searching.Second,through the study of existing filtering algorithm and the research of Filtering algorithm based on Gaussian filter,for better performance of joint bilateral filtering has been improved on speed and filtering effect.Algorithm complexity has been significantly optimized.Speed, through the Thiele rational approximation.By setting the threshold for the color image color,accurately test the edge detection.Different areas to filter segment.Finally,through the experiment,using the adaptive point cloud density neighborhood search algorithm to the original point cloud data.By adjusting the search window, finally get the best search window.Herein the filter algorithm,I conducted a selection experiment parameters,and a selection of the best color threshold.Eventually the filtering effect and filtering speed compare to the original joint bilateral filtering.Compared to the original joint bilateral filtering algorithm the filtering algorithm is faster and better.