基于改进DBSCAN算法的ICESat-2海面点数据去噪处理及精度评估

De-noising and accuracy evaluation of ICESAT-2 sea surface data based on DBSCAN algorithm

ICESat-2卫星搭载先进的光子计数式激光雷达ATLAS,使用全新的测量方式获取海量的光子点云数据。光子计数激光点云由于受海洋表面噪声以及水中穿透性较强等因素影响,现有的去噪算法在浅海区域海面去噪上存在一定的局限性。针对上述问题,本文提出了一种基于改进DBSCAN参数的去噪算法,通过利用点云数据集自身分布特性生成候选Eps参数,自动寻找最优Eps参数,改进现有的DBSCAN去噪算法,并在南海两处岛礁区域进行实验验证。实验表明该算法在浅海区域的去噪精度达到98%,优于传统算法的去噪精度,该研究可为ICESat-2海面点的去噪工作提供参考。

The ICESat-2 satellite is equipped with the advanced photon counting lidar ATLAS,which uses a new measure-ment method to obtain massive photon point cloud data.The photon counting laser point cloud is affected by ocean surface noise and strong water penetration.The existing de-noising algorithms have certain limitations in de-noising the surface of shallow sea areas.In response to the above problems,this paper proposes a de-noising algorithm based on improved DB-SCAN parameters.Using the distribution characteristics of the point cloud data set to generate candidate Eps parameters,and then the Eps parameters are automatically found as the optimal parameters.Experiments were also carried out in two islands and reefs in the South China Sea.The experiment shows that the algorithm has a de-noising accuracy of 98%in the shallow sea area,which is better than the traditional algorithm.This research can provide a reference for the de-noising work of ICE-Sat-2 sea surface points.