Research On Stripe Image Processing For Three Dimensional Laser Mapping

Airborne Lidar has been used in the fields of geological exploration and terrain reconnaissance,etc. In recent years,Airborne Lidar makes the breakthrough to achieve greater distances of high altitude detection. Streak Tube Imaging lidar(STIL) can meet the needs of high altitude detection, It’s echo signal is high frame frequency stripe images. However, because of the high altitude and the influence of the noise of probe and background light source, the stripe images of echo signal is so weak to generate complete and clear 3D point cloud images. So, this paper studies the stripe images based on airborne lidar to generate complete and cleat 3D point cloud images.Based om the gray value statistics of the stripe image, the statistical characteristic of noise, and the different spatial distribution characteristics between the stripe signal and nose, studying denoising algorithm for stripe image using basic spatial domain and frequency domain algorithm. Comparing the SNR(signal to noise ratio) of each algorithm, the SNR of Median filtering algorithm is 20.90 d B, it is superior than others. To get greater SNR, this paper put forward the improved Median filtering algorithm, the SNR reached to 32.66 d B.This paper obtained the denoising algorithm based on stripe image.This paper studies relativity between the adjacent sequence of stripe image. Firstly, got the distribution range of correlation coefficient based on different geographic information such as plain, roads, houses and aircrafts, through doing correlation statistics for 1second stripe image. That provides reference for the selection of the correlation coefficient threshold. Than, this paper makes the analysis for the intensity distribution of the adjacent stripe image, including Intensity distribution characteristics of the double image and distribution characteristics of single stripe signal. Based on the above two characteristics, this paper choses the sequence of stripe images with weak signal, using the methods to make supplement for the parts of weak signal. In this way, this paper got the more compete and clearer 3D point cloud image than the traditional, in the same area, The number of point cloud increases by 45.98%.This paper studies and obtains the best processing method of stripe image through two directions between filtering background noise and enhancing the weak echo stripes, and point the guide for stripe image processing of 3 d laser mapping. The research results obtained puts forward effective solution for the weak echo signal. This paper provides stripe image processing research direction for the breakthrough in more high altitude detection in the future.