A novel denoising algorithm for photon-counting laser data based on LDBSCAN

In this paper, a new single photon laser data processing method is proposed and coarse -to-fine denoising strategy is adopted. Global and piecewise denoising based on the frequency histogram of photon elevation is the first step and direction self-adaptive fine denoising is the next which calculates density value, and uses self-adaptive elliptical LDBSCAN (A Local-Density Based Spatial Clustering Algorithm with Noise) algorithm to effectively remove the noise distributed around the signal segment. Experiments using MABEL (Multiple Altimeter Beam Experimental Lidar) data is implemented and the results validate the proposed algorithm which can effectively extract signal photons from high background noise, and has more reliable results than MABEL official to some extent.