摘要
提出一种利用LiDAR数据进行单株木识别的方法,首先利用广义高斯模型分解全波形LiDAR数据,得到高密度的点云和相应的波形参数,通过建立数字高层模型得到非地面点云,然后计算点云的空间特征得到林木点云,最后在3D空间中利用马尔可夫随机场重新标记得到单株木点云。实验表明,与传统方法相比,本文方法能有效提高单株木识别的准确性,特别是对茂密林地中低矮、细小林木识别效果明显,平均识别精度达到75%。
By analyzing the shortage of traditional approach,a new individual trees recognition method was proposed.Firstly,the generalized Gaussian function was used to analyze the fitting pulse shape LiDAR data,and the high density point cloud and the waveform parameters were obtained,then the non-ground points were gained by establishing DEM;secondly,the spatial characteristics of point cloud was computed to receive forest points;lastly,Markov random fields were exploited to label individual trees in 3D.The experimental results show that this method can effectively improve the recognition accuracy,especially in the low dense,small trees identification effect,and the average recognition accuracy is 75%.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2011年第7期200-203,209,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家高技术研究发展计划(863计划)资助项目(2007AA120501)
国家重点基础研究发展计划(973计划)资助项目(2006CB701300)
关键词
单株木
模式识别
机载激光雷达
马尔可夫随机场
Individual trees
Pattern recognition
Airborne LiDAR
Markov random fields
