The Compression And Recovery Of The 3D Body Point Clouds Based On Compressed Sensing

With the development and progress of computer technology and electronic technology, the wireless body area network has become the hot pot of research. Wireless body area network is a kind of sensor network composed of tiny sensors which is set up in the body to do some physiological data measurement,the technology meet the requirement of the people's growing healthcare. Wireless body area network request of the sensor to erect on the human body model, the human body is the main part in the study of wireless body area network. Due to security and convenient considerations, 3D body point cloud model can completely replace the real human body model to do the related research. However, with the advanced scanning equipment appear constantly and the growing precision of the scanning, the number of point cloud model data become more and more bigger, which increased the difficulty of data storage and transmission.In recent years, Compressed sensing as a kind of new signal sampling theory, ca used a great interest in science. Compressed sensing theory pointed out that if the signal is sparse or sparse in a transform domain, you can through the nonlinear sampling method for compression, and then use the convex optimization algorithm for high-precision reconstruction. Compressed sensing breaks the shackles of traditional sampling law and brings a new revolution to the signal collec tion and compression domain, it makes the compressible progress and the sampling process at the same time and completes the signal to the transformation of information.This paper firstly expounds the concept of 3 D body point cloud as well as the collection and pretreatment process. Because it is very important for subsequent downsizing and restructuring process to establish effective topology relationship, which will affect the algorithm implementation time, we continue to study several algorithms to establish the topological relationship and discuss the advantages and disadvantages in this paper. Discussing the commonly used method of point cloud simplification and surface reconstruction to simplify.Then use the method of block sparse to make point cloud sparse, make it meet the basic requirement of compressed sensing. Finally using orthogonal matching pursuit algorithm to reconstruct the human point cloud after compressed sampling of gaussian random matrix, compared with the original model, the restore result is better.