草地学报 ›› 2015, Vol. 23 ›› Issue (5): 1093-1100.DOI: 10.11733/j.issn.1007-0435.2015.05.029

• 技术研发 • 上一篇    下一篇

高分一号卫星影像特征及其在草地监测中的应用

王磊1,2, 耿君1, 杨冉冉1, 田庆久1, 杨闫君1, 周洋1   

  1. 1. 南京大学国际地球系统科学研究所, 江苏 南京 210093;
    2. 宁夏大学西北退化生态系统恢复与重建教育部重点实验室, 宁夏 银川 750021
  • 收稿日期:2014-09-02 修回日期:2014-12-11 出版日期:2015-10-15 发布日期:2015-12-01
  • 通讯作者: 田庆久
  • 作者简介:王磊(1980-),男,宁夏银川人,博士,讲师,主要从事植被与生态遥感研究,E-mail:WL8999@163.com
  • 基金资助:

    国家科技重大专项(30-Y20A01-9003-12/13);全球变化研究国家重大科学研究计划项目课题(2010CB951503);国家重点基础研究发展计划(973计划)课题(2012CB723206);国家"十二五"科技支撑计划课题(2011BAC07B03)资助

Characteristics and Application of GF-1 Image in Grassland Monitoring

WANG Lei1,2, GENG Jun1, YANG Ran-ran1, TIAN Qing-jiu1, YANG Yan-jun1, ZHOU Yang1   

  1. 1. International Institute for Earth System Science, Nanjing University, Nanjing, Jiangsu Province 210093, China;
    2. Key Laboratory for Restoration and Reconstruction of Degraded Ecosystem in North-western China of Ministry of Education, Ningxia University, Yinchuan, Ningxia 750021, China
  • Received:2014-09-02 Revised:2014-12-11 Online:2015-10-15 Published:2015-12-01

摘要:

为评价高分一号卫星数据的草地监测能力,在分析传感器波段设置、辐射分辨率和光谱响应系数等特征的基础上,以草地为研究对象,提取草地分布信息,计算不同植被指数,结合地面同步观测的草地光谱、地上生物量、覆盖度和叶面积指数等实测数据,通过R2和均方根误差筛选并建立最优估算模型。结果表明:波段设置与部分常用传感器保持了较好的一致性;空间分辨率的提高,增强了地物类型的识别能力,辐射分辨率的提高,增强了数据的层次性;光谱响应系数较好的涵盖了不同草地类型的光谱曲线特征;叶面积指数和生物量的最佳估算模型均为基于比值植被指数的三次多项式模型,覆盖度最佳估算模型为基于归一化植被指数的幂函数模型,并得到了较好的制图效果。

关键词: 高分一号, 植被指数, 草地监测, 遥感估算, 影像特征

Abstract:

In order to evaluate the monitoring ability of GF-1 image in grassland, on the basis of analysis of the characteristics of the band setting, radiometric and spectral response coefficient of sensor, the distributed information was extracted, and the vegetation index of grassland was calculated. With the combination of field-measured spectrum, vegetation coverage, leaf area index and aboveground biomass data, the best vegetation index for grassland parameters was estimated. The optimal model was determined according to R2 and RMSE(root-mean-square error). The results showed that GF-1 sensors kept consistency in band set comparing with other sensors. The improvement of spatial resolution enhanced the identification ability of object types, and the improvement of radiation resolution enhanced the levels of data. The spectral response coefficients covered better the spectral curves of different types of grassland. The correlation of different grassland vegetation parameters and GF-1 vegetation index reached a high level, and met the needs of remote sensing estimation or inversion. The regression analyses showed that the best estimation model for LAI and the biomass of the grassland were cubic polynomial regression model based on RVI(ratio vegetation index), and the best estimation model for the vegetation coverage of the grassland were power function model based on NDVI(normalized difference vegetation index), and the good mapping effect of the research region was obtained.

Key words: GF-1 sensor, Vegetation index, Grassland monitoring, Remote sensing estimation, Image characteristics

中图分类号: