Research On Indoor 3D Scenery High Quality Reconstruction Base On 2D LiDar

There are high requirements of the efficiency and safety at industry scenery which always belongs to indoor operation,so not only the important research meaning,but also the important economy and social value with auto operating technology of industry vehicles.Three dimensions reconstruction is the critical issue of auto operating with industry vehicles,due to the complicated,random and flexible environment which they exist.According to the research on LiDAR systems be widely applied at reality activities,more accurate and convenient three dimensions reconstruction becomes accessible.Therefore,this project tends to conquer the problems of occlusion phenomenon and unbalance sampling with common LiDAR,explores technology of high quality three dimensions reconstruction,researches on adaptive scanning algorithm,develops theory and method of three dimensions construction with LiDAR.First of all,be aimed at conquer the problems of occlusion phenomenon due to single viewpoint emitting laser beams and unbalance sampling at different distance objects due to fixed angular scanning step of common LiDAR,designs a novel three dimensions LiDAR system which is consist with a two dimensions LiDAR,a long arm bracket and a gimble,with anti-occlusion performance and adjustable scanning step.For eliminating the system errors of this novel three dimensions LiDAR system,providing a clever system calibration method base on the feature of scanning twist at the same object within each scanning cycle.Common LiDAR scanning pattern is always fixed angular step scanning,leads to sampling sparsity at further distance objects,thus point clouds registration accurate will descend due to small portion of overlapping,moreover small objects will be missed at long distance;on the other hand,when it meet requirements at long distance scanning,the scanning points will become too dense,leads to data redundancy which need to be compressed.To conquer the problem of unbalance sampling,a adaptive scanning algorithm be provided which base on kernel density estimation to automatically adjust angular scanning step according to expected space resolution.At the end,lots of experiments be carried out to comparison between common LiDAR system and our LiDAR system,experimental results proves that our LiDAR system has excellent anti-occlusion performance in three dimensions reconstruction task and object tracking task,verifies the object detecting function and adjusts scanning angular step according to the object distance with the adaptive scanning algorithm,and its scanning measurement accurate surpasses the common LiDAR system.So that,this three dimensions LiDAR can achieve high quality and high accurate three dimensions scenery reconstruction.