The Effectiveness of Airborne Lidar in The Evaluation of Denoising Algorithm for Spaceborne Photon-counting data

Denoising algorithm is a key step in the processing of spaceborne photon-counting data. With the availability of ICESat-2 (Ice, Cloud and land Elevation Satellite) data, the performance of many denoising algorithms could be evaluated on real data. However, due to the randomness of the raw photoncounting data and diverse terrain, it is not easy to find ground truth theoretically. Approximate reference such as airborne lidar data has the advantages of wide measuring range, high precision and high density, which is a potential ground truth for the denoising algorithms. However, the evaluation error based on airborne lidar data was rarely analyzed. In this paper, 1) the evaluation errors were analyzed theoretically and experimentally based on simulation data with F-value evaluation criteria. The experimental results show that absolute error may be huge with the increase of background noise. Nevertheless, the relative error between algorithms might be in an acceptable range. 2) We analyzed the errors and gave a method to evaluate the error range. Based on the proposed evaluation method, the estimated error range based on airborne LiDAR can be estimated, which may be useful for the performance evaluation of denoising algorithm for spaceborne photon-counting data.