Processing And Application Of Three Dimendion Laser Data Based On Sheep Body

It has many important practical value that using 3D laser point cloud data to realize the entity 3D reconstruction model.And it becomes a hot research in many application fields,such as measurement of surveying and mapping,graphic image processing and so on.Sheep is one of the main livestocks in Inner Mongolia,and it has very important economic value and research significance.The work based on cloud point data of sheep and completed the three-dimensional reconstruction.Using the three-dimensional laser scanner,it can obtain a set of point set data of the sheep's surface and the point set is our reserch object.Because of many facts,just like the issues of the equipment,the target material or the shaking of our arms in the process of the scanning,the samples may contain some redundant points,called point cloud noise data.The noise data would affect the precision of the reconstruction and causes the sheep distortion.With the development of the research techniques,the accuracy of the three-dimentional equipment is higher and higher.Meanwhile,the data quantity is bigger and bigger.But the bigger data quantity is bound to affect the speed of processing and storage efficiency.Therefore,prior to reconstruct the three-dimensional point cloud data,must be denoising,streamline processing.Thesis bases on the sheep point cloud datas of the three-dimention laser scanning,mainly researches the point cloud denoising,point cloud data simplification and data triangel subdivision three aspects.The main content is as follows:(1)Using the Handyscan 3D handheld laser scanner scans the sheep,and we find the point set samples contain some stray noise data.Therefore,in the text improved the algorithm of k-nearest neighbors and put forward method that combined the k-nearest neighbor denosing and median filtering.And finally,the stray noise data was eliminated from the point group.The improved method about k-nearest neighbors search can improve the data searching efficiency.(2)Utilizing different methods to simplification point cloud data of the sheep,uniform bounding box method,streamline curvature method and regional center of gravity based on octree encoding method.According to the result of the simplied point cloud data comed from different simpliy method,we could get some conclusions.Comprehensive analysis of their advantages and disadvantages find that the regional center of gravity based on octree encoding method meets our requirements that reducing the data quantity and retains the feature points.(3)Introducing the Crust triangle subdivision method,and using method of Crust to triangular mesh point cloud data of sheep and obtains good effect of triangular subdivision.