测绘通报 ›› 2019, Vol. 0 ›› Issue (7): 33-38.doi: 10.13474/j.cnki.11-2246.2019.0214

• 学术研究 • 上一篇    下一篇

GF-2支持下的干旱区稀疏植被区植被覆盖度估算

沈谦1, 朱长明1, 张新2   

  1. 1. 江苏师范大学地理测绘与城乡规划学院, 江苏 徐州 221116;
    2. 中国科学院遥感与数字地球研究所遥感科学国家重点实验室, 北京 100101
  • 收稿日期:2018-10-22 出版日期:2019-07-25 发布日期:2019-07-31
  • 通讯作者: 朱长明。E-mail:zhuchangming@jsnu.edu.cn E-mail:zhuchangming@jsnu.edu.cn
  • 作者简介:沈谦(1992-),男,硕士生,主要从事遥感信息智能提取及生态环境遥感等方面研究。E-mail:shenqian_gis@163.com
  • 基金资助:
    国家重点研发计划(2017YFB0504201);国家自然科学基金(41201460);江苏省研究生创新计划(KYCX17_1691/2)

Fractional vegetation cover estimation in aird and rare vegetation area aided by GF-2 remote sensing data

SHEN Qian1, ZHU Changming1, ZHANG Xin2   

  1. 1. School of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, China;
    2. State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China
  • Received:2018-10-22 Online:2019-07-25 Published:2019-07-31

摘要: 现有像元二分模型MODIS植被覆盖度模型因其形式简单、适用性较强的特点被广泛应用于区域植被覆盖度(FVC)的估算。然而,研究表明在沙漠和低植被覆盖的西部干旱区,从250 m的影像上很难精准地获取NDVIveg(全植被覆盖植被指数)和NDVIsoil(全裸土区植被指数)参数。利用常用的直方图累计法获取模型所需参数NDVIveg和NDVIsoil,估算结果存在普遍高估现象。为此,本文首先引入同期获取的GF-2号卫星数据,从GF-2号影像上提取植被覆盖像元;然后,利用Pixel Aggregate方法重采样至250 m分辨率,获取250 m空间分辨率下纯植被和纯裸土像元;最后,将纯植被和纯裸土像元各自空间位置相对应的MODIS NDVI数据最大值作为模型所需NDVIveg和NDVIsoil参数,实现研究区内植被覆盖度的估算。试验通过与线性回归法、多项式回归法和直方图累计像元二分模型法估算结果进行精度对比,结果表明:利用GF-2影像辅助的像元二分模型,精准地获取了低植被覆盖区NDVIveg和NDVIsoil模型参数,提高了干旱区植被覆盖度的估算精度,并有效地抑制了受稀疏植被影响NDVI在干旱区普遍偏高问题导致的FVC高估的现象。

关键词: 高分二号, MODIS NDVI, 像元二分模型, 植被覆盖度

Abstract: The dimidiate pixel model has been widely used in the estimation of fractional vegetation cover because of its simple form and strong applicability. However, studies have shown that selecting parameters in model from 250 m spatial resolution images is very difficult in arid area with low fractional vegetation cover and desert. The estimation result is generally overestimated by using the parameters in model, NDVIveg and NDVIsoil, acquired by the commonly used histogram accumulation method in the low fractional vegetation cover area with more bare soil and desert. So, this paper firstly used GF-2 high-resolution images of the same time with MODIS to map vegetation cover pixel. Then, Pixel Aggregate method was used to resample vegetation cover map from 0.8 m to 250 m spatial resolution to map pure vegetation cover pixels and pure bare soil pixels with 250 m spatial resolution. The maximum value of MODIS NDVI data corresponding to the spatial position of pure vegetation and pure bare soil pixels was taken as the NDVIveg and NDVIsoil parameters required by the model to estimate the vegetation cover in the study area to estimate fractional vegetation cover. The estimation accuracy of linear regression method, nonlinear regression method, dimidiate pixel model based on histogram accumulation method to get parameters and based on GF-2 images to get parameters were compared.The results of experiment indicate that dimidiate pixel model based on GF-2 images accurately selected the parameters, NDVIveg and NDVIsoil, in the low fractional vegetation cover area. This method improves the estimation accuracy of fractional vegetation cover in the arid area, and suppresses the result overestimation in low fractional vegetation cover area due to the effect of the higher NDVI problem in the sparse vegetation area.

Key words: GF-2, MODIS NDVI, dimidiate pixel model, fractional vegetation cover

中图分类号: