弱光束条件下森林区域光子云去噪算法精度研究

Accuracy of Photon Cloud Noise Filtering Algorithm in Forest Area under Weak Beam Conditions

先进地形激光高度计系统(ATLAS)可为全球森林冠层高度测量提供科学数据,利用ATLAS光子云数据可获取森林冠层高度信息。为探究光子云去噪算法在弱光束条件下森林研究区的去噪效果,采用局部距离统计算法、基于密度的聚类(Density-based spatial clustering of applications with noise,DBSCAN)算法和基于粒子群优化(Particle swarm optimization,PSO)模型的PSO-DBSCAN算法在弱光束条件下的森林区域进行了光子云去噪试验,研究了算法的去噪精度,并分析研究区不同特性对于去噪效果的影响。结果表明:PSO-DBSCAN算法在弱光束条件下森林区域去噪精度达到了0.95,满足光子云去噪的精度要求,该算法相对局部距离统计算法和DBSCAN算法表现出更好的去噪效果;相对地形坡度和植被覆盖度,太阳高度角会对算法的去噪结果产生更大的影响。

Advanced topographic laser altimeter system( ATLAS) can provide scientific data for global canopy height measurement. However,due to the characteristics of background noise in the photon data,the traditional algorithm does not study for the forest coverage area under weak beam conditions and there was still few photon cloud noise filtering algorithm can evaluate the accuracy of noise filtering under the condition of weak beam in forest research area. In order to quantify the accuracy of the photon cloud noise filtering algorithm in the forest research area under the condition of weak beam,the accuracies of the local distance statistical algorithm,the density-based spatial clustering of applications with noise( DBSCAN) algorithm and the particle swarm optimization( PSO)-DBSCAN algorithm for photon cloud noise filtering experiments in forest areas under weak beam conditions were studied,and the influence of different characteristics on noise filtering results was analyzed. The results were as follows: the results showed that PSO-DBSCAN algorithm had the accuracy of noise filtering as 0. 95 in the forest area under weak beam conditions,which met the accuracy of photon cloud noise filtering requirements,and the algorithm performed better than the local distance statistical algorithm and the DBSCAN algorithm. The solar elevation angle had greater impact on the noise filtering algorithm than the terrain slope and vegetation coverage.