为了研究机载激光雷达(LiDAR)树高提取技术,以山东省泰安市徂徕山林场为实验区,于2005年5月进行了机载LiDAR数据获取和外业测量.通过对LiDAR点云数据的分类处理,分别得到了试验区的地面点云子集、植被点云子集和高程归一化的植被点云子集.基于高程归一化的植被点云子集计算了上四分位数处的高度,与实地测量的数据进行了比较,并结合中国森林调查规程进行了实用性分析.结果表明:对于较低密度的点云数据,使用分位数法可以较好地进行林分平均高的估计;机载激光雷达技术对树高估计是可行的,精度都高于87%,总体平均精度为90.59%,其中阔叶树的精度高于针叶树.该试验精度可以满足中国二类森林调查规程中平均树高因子的一般商品林和生态公益林的精度要求,对国有商品林小班的调查精度要求(5%)存在一点差距,需要在国有商品林区进一步开展验证工作.对本试验区而言,已经可以满足其作为森林公园生态公益林的调查要求.

During the past decade,there have been dramatic improvements in LiDAR technology and it has been used sucessfully in forestry.To evaluate LiDAR height estimation performance in China,the Culaishan Forest Farm was selected as test site,which is located in Taian,Shandong province.The airborne discrete return LiDAR data were collected on May 13,2005,using Riegl LMS-Q280i laser scanner together with IMU and DGPS.The relative flight height is 800m and the scan angle is 30 degrees.The Laser beam divergence is 0.5mrad.Only first returns were recorded.The LiDAR point density is about 0.35 point per square meter.The tin filter was used to classify LiDAR point cloud data.The ground point dataset,vegetation point dataset and elevation normalized vegetation point dataset were generated for further analysis.The airphotos acquired simultaneously were used as reference during filter parameters selection.Then grid local maximum,upper-quartile and average height were calculated from elevation normalized vegetation point and compared with ground measurements collected in April,2006. The results demonstrated that it is feasible to use airborne LiDAR technology to estimate forest height.As only low density data was available at the test site,quartiles allowed for good tree height estimation in the low LiDAR point density case.The accuracies from all plots were higher than 87% and the total average accuracy was 90.59%.The accuracy of deciduous forest stand was higher than that of coniferous forest stand as deciduous had more flat shape canopy and higher reflectance.This accuracy met the tree height accuracy requirement of general cash forest and ecological forest in the Forest Management Inventory of China. The accuracy was a little bit lower than the requirement of national cash forest(5%) and more validation should be done for this kind of forests.As for this experiment site,the accuracy showed good fulfillment with the surveying requirement of ecological forest in forest parks.