基于随机森林的光子计数激光雷达点云滤波
|
摘要
新一代星载激光雷达卫星ICESat-2首次采用了微脉冲光子计数激光雷达技术,由于单光子探测的灵敏性导致数据在大气和地表下层产生了大量噪声,因此对光子计数激光雷达点云数据实现信号和噪声的分离是开展进一步应用研究的前提和基础。本文选择美国俄勒冈州和弗吉尼亚州2个研究区,采用MATLAS数据,根据光子点云数据的特点构造了12个光子点云特征,对所构造的特征利用随机森林进行变量筛选,用机器学习方法对光子点云进行分类,并将建立好的模型推广到整个研究区。研究结果表明,本文构建的分类器分类总精度达到了96.79%,Kappa系数为0.94,平均生产者精度和用户精度分别为97.1%和96.8%。在相对弱噪声、平坦地形区域和强噪声、复杂地形区域都取得较好的分类结果。本文结果显示了基于少量样本通过机器学习的方法构建模型,可以推广到较大范围区域的光子点云分类应用中。
The new generation of spaceborne laser satellite ICESat-2(the Ice, Cloud, and land Elevation Satellite-2) of NASA(National Aeronautics and Space Administration) has adopted a newly designed micropulse photon counting system, which is the very first time that this technology gets applied in the space environment. Thanks to the high sensitivity of single photon detection technology, it can be seen from the currently released data product(both from the airborne simulators and the simulation data) that there is huge noise in the atmosphere and even below the ground. Therefore, preliminary research on these relevant experimental data to investigate the methods for separating signal photons from noise photons are important for the future applications.MATLAS data, which simulate the expected performance of the ICESat-2 ATLAS(Advanced Topographic Laser Altimeter System) instrument, was chosen to test our machine learning-based approach from two test sites in Oregon and Virginia in the United States. We first derived 12 features, such as the kNN(k-Nearest Neighbour)distance, based on the characteristics of photon point clouds data. Then we applied feature selection techniques by ranking variable importance using Random Forest. Three most representative features were chosen according to the variable importance ranking and we built a Random Forest classifier trained by the sample points we had selected. The established models were further applied to the whole study area. The final classification results indicate that the classifier we constructed had good performance to distinguish signal photons from noise photons. In terms of the mean values of the statistical indicators in the test sites, the overall classification accuracy was 96.79%, and the Kappa coefficient was 0.94. The producer and user accuracies were 97.1% and96.8%, respectively. Additionally, the results show that our method not only worked well on data of relatively lower noise rate on flat terrain surfaces but also achieved good results for those with higher noise rate on complex terrain surfaces. To conclude, our method showes good potential to be applied to larger areas, for especially the classification of the photon counting LiDAR data in the future.
The new generation of spaceborne laser satellite ICESat-2(the Ice, Cloud, and land Elevation Satellite-2) of NASA(National Aeronautics and Space Administration) has adopted a newly designed micropulse photon counting system, which is the very first time that this technology gets applied in the space environment. Thanks to the high sensitivity of single photon detection technology, it can be seen from the currently released data product(both from the airborne simulators and the simulation data) that there is huge noise in the atmosphere and even below the ground. Therefore, preliminary research on these relevant experimental data to investigate the methods for separating signal photons from noise photons are important for the future applications.MATLAS data, which simulate the expected performance of the ICESat-2 ATLAS(Advanced Topographic Laser Altimeter System) instrument, was chosen to test our machine learning-based approach from two test sites in Oregon and Virginia in the United States. We first derived 12 features, such as the kNN(k-Nearest Neighbour)distance, based on the characteristics of photon point clouds data. Then we applied feature selection techniques by ranking variable importance using Random Forest. Three most representative features were chosen according to the variable importance ranking and we built a Random Forest classifier trained by the sample points we had selected. The established models were further applied to the whole study area. The final classification results indicate that the classifier we constructed had good performance to distinguish signal photons from noise photons. In terms of the mean values of the statistical indicators in the test sites, the overall classification accuracy was 96.79%, and the Kappa coefficient was 0.94. The producer and user accuracies were 97.1% and96.8%, respectively. Additionally, the results show that our method not only worked well on data of relatively lower noise rate on flat terrain surfaces but also achieved good results for those with higher noise rate on complex terrain surfaces. To conclude, our method showes good potential to be applied to larger areas, for especially the classification of the photon counting LiDAR data in the future.
引文
[1]庞勇,李增元,陈尔学,等.激光雷达技术及其在林业上的应用[J].林业科学,2005,41(3):129-136.[Pang Y, Li Z Y,Chen E X, et al. Lidar remote sensing technology and its application in forestry[J]. Scientia Silvae Sinicae, 2005,41(3):129-136.]
[2] Rosette J, Cook B, Nelson R, et al. Sensor compatibility for biomass change estimation using remote sensing data sets:Part of NASA's carbon monitoring system initiative[J]. IEEE Geoscience and Remote Sensing Letters, 2015,12(7):1511-1515.
[3]付甜,庞勇,黄庆丰,等.亚热带森林参数的机载激光雷达估测[J].遥感学报,2011,15(5):1092-1104.[Fu T, Pang Y,Huang Q F, et al. Prediction of subtropical forest parameters using airborne laser scanner[J]. Journal of Remote Sensing, 2011,15(5):1092-1104.]
[4]黄克标,庞勇,舒清态,等.基于ICESatGLAS的云南省森林地上生物量反演[J].遥感学报,2013,17(1):65-179.[Huang K B, Pang Y, Shu Q T, et al. Aboveground forest biomass estimation using ICESat GLAS in Yunnan, China[J]. Journal of Remote Sensing, 2013,17(1):65-179.]
[5]徐新良,曹明奎.森林生物量遥感估算与应用分析[J].地球信息科学学报,2006,8(4):122-128.[Xu X L, Cao M K. An analysis of the applications of remote sensing method to the forest biomass estimation[J]. Journal of Geo-information Science, 2006,8(4):122-128.]
[6]刘洋,刘荣高,陈镜明,等.叶面积指数遥感反演研究进展与展望[J].地球信息科学学报,2013,15(5):734-743.[Liu X, Liu R G, Chen J M, et al. Current status and perspectives of leaf area index retrieval from optical remote sensing data[J]. Journal of Geo-information Science, 2013,15(5):734-743.]
[7] Los S O, Rosette J, Kljun N, et al. Vegetation height products between 60°S and 60°N from ICESat GLAS data[J]. Geoscientific Model Development, 2012,5:413-432.
[8] Markus T, Neumann T, Martino A, et al. The Ice, Cloud,and land Elevation Satellite-2(ICESat-2):Science requirements, concept, and implementation[J]. Remote Sensing of Environment, 2017,190:260-273.
[9] Neuenschwander A L, Magruder L A. The potential impact of vertical sampling uncertainty on ICESat-2/ATLAS terrain and canopy height retrievals for multiple ecosystems[J]. Remote Sensing, 2016,8(12):1039.
[10] McGill M, Markus T, Scott V S, et al. The multiple altimeter beam experimental Lidar(MABEL):An airborne simulator for the ICESat-2 mission[J]. Journal of Atmospheric and Oceanic Technology, 2013,30(2):345-352.
[11] Brown M E, Arias S D, Neumann T, et al. Applications for ICESat-2 data:From NASA's early adopter program[J]. IEEE Geoscience and Remote Sensing Magazine,2016,4(4):24-37.
[12]夏少波,王成,习晓环,等. ICESat-2机载试验点云滤波及植被高度反演[J].遥感学报,2014,18(6):1199-1207.[Xia S B, Wang C, Xi X H, et al. Point cloud filtering and tree height estimation using airborne experiment data of ICESat-2[J]. Journal of Remote Sensing, 2014,18(6):1199-1207.]
[13] Awadallah M, Ghannam S, Abbott A L, et al. Active contour models for extracting ground and forest canopy curves from discrete laser altimeter data[C]//Proceedings:13th International Conference on LiDAR Applications for Assessing Forest Ecosystems, 2013:129-136.
[14] Herzfeld U C, McDonald B W, Wallin B F, et al. Algorithm for detection of ground and canopy cover in micropulse photon-counting Lidar altimeter data in preparation for the ICESat-2 mission[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52:2109-2125.
[15] Zhang J, Kerekes J. An adaptive density-based model for extracting surface returns from photon-counting laser altimeter data[J]. IEEE Geoscience and Remote Sensing Letters, 2015,12(4):726-730.
[16] Gwenzi D, Lefsky M A, Suchdeo V P, et al. Prospects of the ICESat-2 laser altimetry mission for savanna ecosystem structural studies based on airborne simulation data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016,118:68-82.
[17] Nie S, Wang C, Xi X, et al. Estimating the vegetation canopy height using micro-pulse photon-counting LiDAR data[J]. Optics express, 2018,26(10):A520-A540.
[18]张继贤,林祥国,梁欣廉.点云信息提取研究进展和展望[J].测绘学报,2017,46(10):1460-1469.[Zang J X, Lin X G, Liang X L. Advances and prospects of information extraction from point clouds[J]. Acta Geodaetica et Cartographica Sinica, 2017,46(10):1460-1469.]
[19] Chen B, Pang Y, Li Z, et al. Ground and top of canopy extraction from photon counting LiDAR data using local outlier factor with ellipse searching area[J]. IEEE Geoscience and Remote Sensing Letters, in press, doi:10.1109/LGRS.2019.2899011.
[2] Rosette J, Cook B, Nelson R, et al. Sensor compatibility for biomass change estimation using remote sensing data sets:Part of NASA's carbon monitoring system initiative[J]. IEEE Geoscience and Remote Sensing Letters, 2015,12(7):1511-1515.
[3]付甜,庞勇,黄庆丰,等.亚热带森林参数的机载激光雷达估测[J].遥感学报,2011,15(5):1092-1104.[Fu T, Pang Y,Huang Q F, et al. Prediction of subtropical forest parameters using airborne laser scanner[J]. Journal of Remote Sensing, 2011,15(5):1092-1104.]
[4]黄克标,庞勇,舒清态,等.基于ICESatGLAS的云南省森林地上生物量反演[J].遥感学报,2013,17(1):65-179.[Huang K B, Pang Y, Shu Q T, et al. Aboveground forest biomass estimation using ICESat GLAS in Yunnan, China[J]. Journal of Remote Sensing, 2013,17(1):65-179.]
[5]徐新良,曹明奎.森林生物量遥感估算与应用分析[J].地球信息科学学报,2006,8(4):122-128.[Xu X L, Cao M K. An analysis of the applications of remote sensing method to the forest biomass estimation[J]. Journal of Geo-information Science, 2006,8(4):122-128.]
[6]刘洋,刘荣高,陈镜明,等.叶面积指数遥感反演研究进展与展望[J].地球信息科学学报,2013,15(5):734-743.[Liu X, Liu R G, Chen J M, et al. Current status and perspectives of leaf area index retrieval from optical remote sensing data[J]. Journal of Geo-information Science, 2013,15(5):734-743.]
[7] Los S O, Rosette J, Kljun N, et al. Vegetation height products between 60°S and 60°N from ICESat GLAS data[J]. Geoscientific Model Development, 2012,5:413-432.
[8] Markus T, Neumann T, Martino A, et al. The Ice, Cloud,and land Elevation Satellite-2(ICESat-2):Science requirements, concept, and implementation[J]. Remote Sensing of Environment, 2017,190:260-273.
[9] Neuenschwander A L, Magruder L A. The potential impact of vertical sampling uncertainty on ICESat-2/ATLAS terrain and canopy height retrievals for multiple ecosystems[J]. Remote Sensing, 2016,8(12):1039.
[10] McGill M, Markus T, Scott V S, et al. The multiple altimeter beam experimental Lidar(MABEL):An airborne simulator for the ICESat-2 mission[J]. Journal of Atmospheric and Oceanic Technology, 2013,30(2):345-352.
[11] Brown M E, Arias S D, Neumann T, et al. Applications for ICESat-2 data:From NASA's early adopter program[J]. IEEE Geoscience and Remote Sensing Magazine,2016,4(4):24-37.
[12]夏少波,王成,习晓环,等. ICESat-2机载试验点云滤波及植被高度反演[J].遥感学报,2014,18(6):1199-1207.[Xia S B, Wang C, Xi X H, et al. Point cloud filtering and tree height estimation using airborne experiment data of ICESat-2[J]. Journal of Remote Sensing, 2014,18(6):1199-1207.]
[13] Awadallah M, Ghannam S, Abbott A L, et al. Active contour models for extracting ground and forest canopy curves from discrete laser altimeter data[C]//Proceedings:13th International Conference on LiDAR Applications for Assessing Forest Ecosystems, 2013:129-136.
[14] Herzfeld U C, McDonald B W, Wallin B F, et al. Algorithm for detection of ground and canopy cover in micropulse photon-counting Lidar altimeter data in preparation for the ICESat-2 mission[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52:2109-2125.
[15] Zhang J, Kerekes J. An adaptive density-based model for extracting surface returns from photon-counting laser altimeter data[J]. IEEE Geoscience and Remote Sensing Letters, 2015,12(4):726-730.
[16] Gwenzi D, Lefsky M A, Suchdeo V P, et al. Prospects of the ICESat-2 laser altimetry mission for savanna ecosystem structural studies based on airborne simulation data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016,118:68-82.
[17] Nie S, Wang C, Xi X, et al. Estimating the vegetation canopy height using micro-pulse photon-counting LiDAR data[J]. Optics express, 2018,26(10):A520-A540.
[18]张继贤,林祥国,梁欣廉.点云信息提取研究进展和展望[J].测绘学报,2017,46(10):1460-1469.[Zang J X, Lin X G, Liang X L. Advances and prospects of information extraction from point clouds[J]. Acta Geodaetica et Cartographica Sinica, 2017,46(10):1460-1469.]
[19] Chen B, Pang Y, Li Z, et al. Ground and top of canopy extraction from photon counting LiDAR data using local outlier factor with ellipse searching area[J]. IEEE Geoscience and Remote Sensing Letters, in press, doi:10.1109/LGRS.2019.2899011.