测绘通报 ›› 2019, Vol. 0 ›› Issue (10): 17-22.doi: 10.13474/j.cnki.11-2246.2019.0311

• 自然资源监测 • 上一篇    下一篇

草原植被覆盖度遥感估算模型的适用性比较

董显聪1,2, 李晓洁1   

  1. 1. 中国科学院东北地理与农业生态研究所, 吉林 长春 130012;
    2. 中国科学院大学, 北京 100049
  • 收稿日期:2019-06-25 出版日期:2019-10-25 发布日期:2019-10-26
  • 通讯作者: 李晓洁。E-mail:lixiaojie@iga.ac.cn E-mail:lixiaojie@iga.ac.cn
  • 作者简介:董显聪(1997-),男,硕士生,主要研究方向为基于深度学习的遥感运算。E-mail:dongxiancong@126.com
  • 基金资助:
    国家自然科学基金面上项目(41671350)

Comparison of applicability of remote sensing estimation model for grassland vegetation coverage

DONG Xiancong1,2, LI Xiaojie1   

  1. 1. Institute of Northeast Geography and Agricultural Ecology, Chinese Academy of Sciences, Changchun 130012, China;
    2. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2019-06-25 Online:2019-10-25 Published:2019-10-26

摘要: 植被覆盖度及其变化对区域生态系统的稳定性具有直接影响,且这种影响在草原地区更加明显。为探寻草原植被覆盖度的最佳遥感估算方法,本文对像元二分模型、Carlson模型和Baret模型的估算精度和适用性进行了比较,优化了Baret模型的参数,以提高其在草原地区的估算精度。内蒙古呼伦贝尔地区的草地计算结果表明:像元二分模型有高估植被覆盖度的现象;Carlson模型在低植被覆盖区低估了植被覆盖度,而在高植被覆盖区高估了植被覆盖度;Baret模型在草原地区的估算精度最高。对Baret模型进行参数优化后,其在高植被覆盖度区域的估算精度提升了4.9%。

关键词: 草原植被覆盖度, 照相法, NDVI, 混合像元模型

Abstract: Vegetation coverage and its changes have a direct impact on the stability of regional ecosystem, and this kind of effect is more obvious in grassland areas. In order to explore the best remote sensing estimation method of grassland vegetation coverage, this paper compares the estimation accuracy and applicability of binary pixel model, Carlson model and Baret model, and finally optimizes the parameters of Baret model to improve its estimation accuracy in grassland areas. The results of grassland calculation in Hulun Buir region of Inner Mongolia show that pixel dichotomy model overestimates the vegetation coverage, while Carlson model underestimates the vegetation coverage in low vegetation-covered areas and overestimates it in high vegetation-covered areas. And Baret model has the highest estimation accuracy in grassland areas. After parameter optimization of Baret model, the estimation accuracy in high vegetation-covered areas is improved by 4.9%.

Key words: grassland vegetation coverage, photographic method, NDVI, mixed pixel model

中图分类号: