点云密度对激光雷达估计森林样方平均树高的影响  

Effects of LiDAR Point Density on Tree Height Estimation in Plots Level

作　　者：尤号田[1] 邢艳秋[1] 王铮[2] 王蕊[1] 孙小添[1] 

机构地区：[1]东北林业大学,哈尔滨150040 [2]吉林省基础地理信息中心

出　　处：《东北林业大学学报》2014年第5期143-148,共6页Journal of Northeast Forestry University

基　　金：中央高校基本科研业务费专项资金支撑项目(DL12EB07);国家自然科学基金支撑项目(41171274)资助

摘　　要：以吉林省长春市净月潭国家森林公园为研究区,分别于2012年5月和10月进行飞行数据和野外数据采集。首先通过对小光斑激光雷达离散点云数据进行随机稀释操作,获得四种不同点云密度数据,再对点云数据进行分层处理并拟合伪波形,从波形中提取冠层半能量高用于估测森林的样方平均树高。结果表明:在研究的点云密度范围内无论点云密度的高低,冠层半能量高模型均能够较好的估测森林样方平均树高;四种点云密度情况下,0.125倍点云密度时,模型结果相对较好,拟合相关性R=0.971,精度P=0.97;不同点云密度对模型拟合相关性及精度的影响差异不大,且落叶松的估测精度高于樟子松。In Changchun Moon Lake National Forest Park, we collected the flight data and field data in May and October 2012, respectively. We first thinned the small footprint LiDAR points by random selecting, and then sliced the data by layers to get four different kinds density data. We fitted the pseudo waveform and estimated the forest plots average tree height extracted from canopy medium energy. All the models within the different density levels can estimate the forest plots average tree height well. The height model with canopy medium energy in 0. 125 times density level is better with the fitting correlation of 0.971 and the accuracy of 0.97. The effect of different density points on the fitting relevance and accuracy of the models is insignificant. The accuracy of larch forest stand is higher than that of pine forest stand.

关 键 词：点云密度 小光斑激光雷达 平均树高 伪波形 冠层半能量高 

分 类 号：S771.8[农业科学—森林工程]