Research On Fast Processing Technology Of Lidar Echo Data

Airborne LiDAR is an important means of geological and topographic survey. After several decades of development, the use and system are more and more abundant. The LiDAR system applying streak tube, which is a new system of Li DAR, is the characteristics of high sensitivity, high frame rate and high resolution. It can be obtained strength and distance information of the target at the same time, and has broad application prospects. With rapid development of the streak tube LiDAR, its data acquisition efficiency has an order of increase. How to efficiently deal with the massive echo data becomes the key technology of the streak tube Li DAR. Therefore, it is very important to research on the fast processing of the echo data.This article describes the airborne LiDAR working principle and echo signal processing, finds the main factors that affect the speed of data processing, compares two software on the merits of the echo signal processing and uses this software for parallel acceleration. Finally, based on the morphological method, the classification and extraction of the target are carried out.For echo signal denoising process is a major factor affecting the data processing speed, MATLAB and LabVIEW two kinds of software operating environment of the echo signal denoising efficiency was compared. Compared to LabVIEW, MATLAB has a faster processing speed. Using the MATLAB software environment, the CPU based parallel processing method is used for the fast processing of the echo signal denoising process. It is proved that using parallel processing technology can significantly improve the processing speed of the echo signal. The maximum speedup can reach 3.0. In order to further improve the efficiency of echo signal processing, the parallel processing technology of GPU is researched in this article. When using CPU and GPU to cooperate with the way to accelerate, the maximum speedup can reach 9.4.In the research of target classification technology, according to the characteristics of the LiDAR echo signal, the morphological method is used to the classification of the original echo data. Through the classification of outfield experimental echo signal, it is proved that the classification of buildings, roads, slopes and other targets is feasible. The classification error rate is controlled at about 10% and the classification speed is about 3800 points per second. It is proved that the classification based on the original echo signal can be realized based on the morphological method in this article.