欢迎访问林草资源研究

林草资源研究 ›› 2024›› Issue (2): 68-79.doi: 10.13466/j.cnki.lczyyj.2024.02.009

• 科学研究 • 上一篇    下一篇

基于无人机影像的高郁闭度马尾松林蓄积量估算方法研究

骆耀培1,2(), 李和平3(), 杨广斌1,2, 岑刚4, 李蔓1,2, 曹乾洋1,2, 王仁儒1,2, 陈盼芳3   

  1. 1.贵州师范大学 地理与环境科学学院,贵阳 550025
    2.贵州省山地资源与环境遥感应用重点实验室,贵阳 550025
    3.贵州省第一测绘院,贵阳 550025
    4.贵州省林业调查规划院,贵阳 550003
  • 收稿日期:2023-12-11 修回日期:2024-01-18 出版日期:2024-04-28 发布日期:2024-09-02
  • 通讯作者: 李和平,高级工程师,主要从事地理信息系统与遥感技术研究。Email:1596686651@qq.com
  • 作者简介:骆耀培,硕士研究生,主要从事地图学与地理信息系统研究。Email:lypcp1874@163.com
  • 基金资助:
    贵州省科技计划项目“基于遥感大数据的自然资源统计和资产价值评估”(黔科合重大专项[2022]001);贵州省科技计划项目“基于喀斯特地区自然资源资产遥感监测关键技术研究”(黔科合支撑[2023]一般176)

Volume Estimation Method of High Canopy Density Pinus massoniana Forest Based on UAV Image

LUO Yaopei1,2(), LI Heping3(), YANG Guangbin1,2, CEN Gang4, LI Man1,2, CAO Qianyang1,2, WANG Renru1,2, CHEN Panfang3   

  1. 1. School of Geography and Environmental Science,Guizhou Normal University,Guiyang 550025,China
    2. Key Laboratory of Mountain Resources and Environmental Remote Sensing Application,Guiyang 550025,China
    3. Guizhou First Surveying and Mapping Institute,Guiyang 550025,China
    4. Forest Survey and Planning Institute of Guizhou Province,Guiyang 550003,China
  • Received:2023-12-11 Revised:2024-01-18 Online:2024-04-28 Published:2024-09-02

摘要:

为降低传统野外调查成本,提高高郁闭度森林资源调查效率,以多光谱无人机影像结合实地样地调查数据为源数据,以马尾松纯林为研究对象,采用冠层高度模型(CHM)及多光谱影像6种植被指数构建模型,估算研究区林分蓄积量。结果表明:1)高分辨率数字高程模型的辅助能够有效弥补无人机影像在茂密森林无法提取地面点的缺陷,可提高CHM构建的精度,实现在茂密森林树高的准确提取;2)采用CHM提取研究区单木树高并估算蓄积量时,样地内共提取马尾松292株,提取的平均树高为18.77 m,小班区域内共提取马尾松18 120株,提取的平均树高为17.02 m,实测平均树高为18.17 m,平均树高提取效果较好,估算蓄积量为7 466.74 m3,实测蓄积量为9 024.40 m3,蓄积量估算精度为82.90%;3)植被指数模型的RMSE为0.39,R2为0.84,模型精度较高,蓄积量估算为8 620.30 m3,估算精度为96.26%。通过借助无人机遥感技术,两种蓄积量估算方法均能在高郁闭度森林中实现蓄积量的快速估算,其中通过多光谱数据提取植被指数构建模型估算蓄积量的效果更佳,这为进一步推广无人机影像在高郁闭度森林资源调查中的应用提供了有力的科学依据。

关键词: 无人机影像, 高郁闭度, 马尾松, 冠层高度模型, 植被指数, 蓄积量

Abstract:

In order to reduce the cost of traditional field investigation and improve the efficiency of high canopy density forest resource investigation,multi-spectral UAV images combined with field sample survey data were used as the source data,and the Pinus massoniana pure stand was used as the research object.The stand stock volume in the study area was estimated by using the canopy height model(CHM)and six vegetation indices from multi-spectral images.The results show that:1)The assistance of a high-resolution digital elevation model can effectively compensate for the defect that UAV images can not extract ground points from in dense forests,improve the accuracy of CHM construction,and achieve accurate extraction of tree height in dense forests.2)When CHM was used to extract the tree height of a single tree in the study area and estimate the volume,292 Pinus massoniana were extracted from the plot,and the average tree height was 18.77 m.A total of 18 120 Pinus massoniana were extracted from the subcompartment area,and the average tree height was 17.02 m.The measured average tree height was 18.17 m.The average tree height extraction effect was good.The estimated volume was 7 466.74 m3,the measured volume was 9 024.40 m3,and the accuracy of the estimation was 82.90%.3)The RMSE of the vegetation index model is 0.39,R2=0.84,and the accuracy of the model is high.The volume is estimated to be 8 620.30 m3,and the estimation accuracy is 96.26%.By using UAV remote sensing technology,both of the two stock volume estimation methods can achieve rapid estimation of stock volume in high-canopy density forests.Among them,the effect of estimating stock volume by extracting vegetation index from multi-spectral data is better.This provides a strong scientific basis for further promotion and application of UAV images in surveys of forest resources with high canopy density.

Key words: UAV image, high canopy density, Pinus massoniana, canopy height model, vegetation index, stock volume

中图分类号: