自然资源学报 ›› 2008, Vol. 23 ›› Issue (3): 507-513.doi: 10.11849/zrzyxb.2008.03.018

• 资源研究方法 • 上一篇    下一篇

基于小光斑激光雷达数据的稀疏林地冠层高程重建

覃驭楚1,3, 吴运超1,3, 牛铮1, 占玉林1, 熊在平2   

  1. 1. 中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101;
    2. 中国科学院沈阳应用生态研究所,沈阳110016;
    3. 中国科学院研究生院,北京100049
  • 收稿日期:2006-10-31 修回日期:2007-10-31 出版日期:2008-05-28 发布日期:2008-05-28
  • 通讯作者: E-mail:zheng_niu@263.net E-mail:zheng_niu@263.net
  • 作者简介:覃驭楚(1982-),湖北宜昌人,硕博连读,研究方向为遥感和地理信息系统应用。E-mail:yuchuqin@hotmail.com
  • 基金资助:

    国家重点基础研究发展规划项目(2007CB714406);地理空间信息工程国家测绘局重点实验室资助项目(200714);国家自然科学基金项目(40701119)

Reconstruction of Sparse Forest Canopy Height Using Small Footprint LiDAR Data

QIN Yu-chu1,3, WU Yun-chao1,3, NIU Zheng1, ZHAN Yu-lin1, XIONG Zai-ping2   

  1. 1. State Key Laboratory of Remote Sensing Science,Institute of Remote Sensing Applications,CAS,Beijing 100101,China;
    2. Institute of Applied Ecology,CAS,Shenyang 110016,China;
    3. Graduate University of Chinese Academy of Sciences,Beijing 100049,China
  • Received:2006-10-31 Revised:2007-10-31 Online:2008-05-28 Published:2008-05-28

摘要: 探讨在稀疏林地的条件下,利用小光斑激光雷达获取的多次回波点云数据估算冠层高度,研究不同的高度重建方法包括反距离权重、样条插值以及普通克里格等方法的冠层和地面高度重建的精度。研究发现,在地面和林冠层,不同的重建方法的表现是不一样的,其中样条方法在冠层重建中的精度较高,误差绝对值的平均值为0.95m,并且整体偏差也较小,方差为3.42。而在地面高程重建则是普通克里格方法具有较小的误差,误差绝对值的平均值为0.35m,而样条方法的整体偏差要优于其他两种方法,方差为0.48。经过综合考虑,利用样条方法重建的冠层高程和普通克里格方法重建的地面高程实现了实验区冠层高度的提取。

关键词: 小光斑激光雷达, 稀疏林地, 冠层高度, 估算

Abstract: This study estimates canopy height using multi-returns data acquired by the small footprint lidar of sparse forest and explores the precision of different methods such as IDW(Inverse distance weight),Spline method,OK(Ordinary Kriging) to reconstruct the canopy elevation and the ground elevation.It is found out that the performance of the different methods is different between the condition of the forest canopy and the ground.Thereinto the Spline method has the best precision on the reconstruction of forest canopy,the mean of absolute value error is 0.95m,and the variance is 3.42.But in the case of ground,the OK method's mean of absolute value error is the lowest,being 0.35m,but the variance of Spline method is the lowest,being 0.48.But integrating the precision of ground elevation with canopy elevation,and choosing the canopy elevation reconstructed by Spline and the ground elevation reconstructed by OK,it is possible to achieve the abstraction of the canopy height of the study area.

Key words: small footprint LiDAR, sparse forest, canopy height, estimate

中图分类号: 

  • S758