Acta Agrestia Sinica ›› 2020, Vol. 28 ›› Issue (6): 1664-1672.DOI: 10.11733/j.issn.1007-0435.2020.06.021

Previous Articles     Next Articles

Research on Vegetation Extraction and Fractional Vegetation Cover of Karst Area Based on Visible Light Image of UAV

YIN Lin-jiang1,2, ZHOU Zhong-fa1,2, LI Shao-hui1,2, HUANG Deng-hong1,2   

  1. 1. Institute of Karst Science/School of Geography and Environmental Science, Guizhou Normal University, Guiyang, Guizhou Province 550001, China;
    2. State Engineering Technology Institute for Karst Desertification Control, Guiyang, Guizhou Province 550001, China
  • Received:2020-06-25 Revised:2020-07-08 Published:2020-12-02

基于无人机可见光影像对喀斯特地区植被信息提取与覆盖度研究

尹林江1,2, 周忠发1,2, 李韶慧1,2, 黄登红1,2   

  1. 1. 贵州师范大学喀斯特研究院/地理与环境科学学院, 贵州 贵阳 550001;
    2. 国家喀斯特石漠化防治工程技术研究中心, 贵州 贵阳 550001
  • 通讯作者: 周忠发
  • 作者简介:尹林江(1993-),男,贵州德江县人,硕士研究生,主要从事无人机山地遥感、GIS开发与应用研究,E-mail:ylj8575@163.com
  • 基金资助:
    国家自然科学基金地区项目(41661088)资助;国家重点研发计划项目(2018YFB0505400)资助;贵州省高层次创新型人才培养计划“百”层次人才(黔科合平台人才〔2016〕5674)资助

Abstract: In order to obtain the vegetation and vegetation coverage information in karst rocky desertification area accurately and quickly,the visible light image of rocky desertification area was acquired by UAV. The excess green (EXG),visible band difference vegetation index (VDVI),red green and blue vegetation index (RGBRI),and excess green-excess red (ExG-ExR) were selected. Based on the idea of intersection method and sample statistics method of vegetation index time series diagram,the threshold segmentation method was used to extract the information of vegetation and vegetation coverage. The vegetation and vegetation coverage information obtained by supervised classification was as the true values and the accuracy was verified. The results showed that ExG-ExR had the highest accuracy for vegetation information extraction in Huajiang karst rocky desertification area of Guanling,Guizhou Province,with the overall accuracy of 95.56% and kappa coefficient of 0.919.The vegetation coverage obtained by ExG-ExR index was the closest to the real value,with RMSE of 0.097 and R2 of 0.977. It indicated that in karst area,using the intersection method of vegetation index time series graph and the idea of sample statistics was suitable and accurate for the extraction of vegetation information and vegetation coverage in this area.

Key words: UAV image, Karst rocky desertification, Vegetation extraction, Fractional vegetation cover, Threshold extraction

摘要: 为准确快速获取喀斯特石漠化地区植被和植被覆盖度信息,本研究利用四旋翼无人机采集喀斯特石漠化区域的可见光影像,选择过绿指数(EXG,excess green)、可见光波段差异植被指数(VDVI,visible-band difference vegetation indx)、红绿蓝植被指数(RGBRI,red green and blue vegetation index)、过绿减过红指数(ExG-ExR,excess green-excess red),利用植被指数时序图交点法和样本统计法思想,运用阈值分割法进行植被和植被覆盖度的信息提取,并以监督分类得到的植被和植被覆盖度信息为真实值,进行精度验证。结果表明,在贵州省关岭贞丰花江喀斯特石漠化地区,对于植被信息提取,ExG-ExR的精度最高,总体精度为95.56%,Kappa系数为0.919;ExG-ExR得到的植被覆盖度精度最好,为99.174%,均方根误差RMSE为0.097,R2为0.977。由此可见,在喀斯特地区利用植被指数时序图交点法和样本统计法思想,适合该地区的植被信息提取和植被覆盖度的提取,具有较高的精度。

关键词: 无人机影像, 喀斯特石漠化, 植被信息, 植被覆盖度, 阈值提取

CLC Number: