Research on Adaptive Denoising Algorithm of Single Photon Laser Altimetry Data for On Orbit Processing

The new generation of laser mapping satellite represented by ICESat-2 adopted multi-beam photon counting detection technology. A large amount of noise was attached to the detection data, which brings challenges to the satellite-ground transmission and subsequent processing. For the application requirement of efficient on orbit processing, a multi-level joint denoising method based on local photon elevation histogram statistical filtering was proposed. First, the obvious noise photons were eliminated by rough surface elevation threshold. Second, elevation histogram statistical filtering and threshold adaptive correction method were used to achieve precise denoising. Comprehensively the method was verified on MABEL laser altimetry data. The verification results indicate that the proposed method has efficient denoising ability and dynamic adaptability, data processing can be carried out under all-day condition of different point cloud density and the average noise elimination rate is above 86% for the daytime’s strong noise under the condition of near lossless retention of effective photons, which realizes efficient denoising of raw data. The use constraints of previous classical algorithms on observation time, photon density, terrain slope, etc have been synchronously improved. Therefor it can be a reference for the design of on-orbit processing algorithm of spaceborne single photon laser altimetry radar.