基于机载激光雷达的林隙结构参数提取  

Canopy Gap Structure Parameters Extraction Based on Light Detection and Ranging(LiDAR)

作　　者：李欢 李明泽[1] 范文义[1] 王斌[1] Li Huan;Li Mingze;Fan Wenyi;Wang Bin(College of Forestry,Northeast Forestry University Harbin 150040)

机构地区：[1]东北林业大学林学院,哈尔滨150040

出　　处：《林业科学》2018年第10期98-107,共10页Scientia Silvae Sinicae

基　　金：国家自然科学基金项目(31470640)

摘　　要：【目的】运用激光雷达技术识别提取林隙面积、边界木高和形状指数,为林隙结构参数调查提供技术支持。【方法】以机载激光雷达数据为数据源,以东北林业大学帽儿山实验林场内设置的5块林隙调查样地中的54个林隙为研究对象,通过对比普通克里金插值、反距离权重插值、样条插值、自然临近插值和局部多项式插值确定最优插值方法得到冠层高度模型,运用交替贯序滤波法对冠层高度模型进行林隙识别提取。【结果】通过对比5种插值方法的RMSE发现,普通克里金法适合插值DEM,反距离权重法适合插值DSM;采用交替贯序滤波法识别提取林隙时,林隙识别率为92.6%,运用配对检验法对提取的林隙面积、边界木高与野外调查数据进行检验,基于激光雷达技术提取的林隙面积和边界木高与野外实测值呈较强线性关系,R^2分别为0.983和0.737,其中提取的林隙面积与实测值平均相对误差为15.78%,林隙边界木高与实测值平均相对误差为11.94%。【结论】基于机载激光雷达数据采用交替贯序滤波能有效识别林隙且能准确提取其结构参数;坡度、坡位及冠层结构复杂程度都会影响冠层高度模型的插值精度;林隙面积提取受林隙边界形状和林层结构影响,林隙边界木高随着样地地形、坡度增加误差也随之增大。机载激光雷达技术具有获取地形和树木三维结构信息的优势,可为林隙识别及其结构参数提取提供新的遥感方法。【Objective】In order to provide technical support for the investigation of gap structure parameters, the LiDAR technology was used in this study to identify the gap structure parameters（including gap size, canopy height and gap shape index）.【Method】 Taking the airborne LiDAR as the data source, the 54 gaps in the five gap investigate plots were set up in the Mao’ershan experimental forest farm of Northeast Forestry University as the study area, The canopy height model was obtained by comparing the five interpolation method （ordinary kriging, inverse distance weight, splines, natural neighbor, local polynomial）to determine the optimal interpolation method. The canopy height model was used to identify and extract the gaps by alternative sequential filter.【Result】 By comparing the RMSE of the five interpolation method , it is found that the ordinary kriging method is suitable for interpolating DEM, and the inverse distance weight method is suitable for interpolating DSM. The gap recognition rate was 92.6% when the gap was extracted by alternating sequential filter. Using the paired test method to check the gap size, canopy height with field survey data, linear correlation was observed between the values of LiDAR estimation and field investigation, and the R 2 values of gap size and canopy height case were 0.983 and 0.737,respectively. Compared with field survey data, the average relative error of extracted gap size was 15.78% and canopy height was 11.94%.【Conclusion】 Using alternating sequential filter based on airborne lidar data can effectively identify the gap and accurately extract its structural parameters. Interpolation accuracy of the canopy height model is affected by the slope, slope position and the complexity of the canopy structure; gap size is affected by boundary shape and forest structure, canopy height as the terrain gradient increases, the error increases. The airborne LiDAR technology has the advantage of acquiring three-dimensional structural information of terrain and tre

关 键 词：机载激光雷达 林隙结构参数 数学形态学 CHM 

分 类 号：S757[农业科学—森林经理学]