北方农业学报 ›› 2011, Vol. ›› Issue (4): 70-70.

• • 上一篇    下一篇

不同遥感影像大气纠正算法在农作物种植面积提取中的对比

王妮[1] 江南[2] 吕恒[2] 彭世揆[1]   

  1. [1]南京林业大学,江苏南京210037 [2]南京师范大学,江苏南京210046
  • 出版日期:2011-08-20 发布日期:2011-08-20
  • 通讯作者: 王妮
  • 作者简介:作者简介:王妮(1984-),女,山东烟台人,南京林业大学博士研究生,主要研究方向为遥感与地理信息系统。 通讯作者:江南(1957-),男,江苏南京人,教授,博士生导师,研究方向为遥感机理与应用研究。
  • 基金资助:
    江苏省普通高校研究生科研创新计划项目(CX09B_188Z); 国家高技术研究发展专项经费资助(2006AA120101)

The Methods of the Atmospheric Correction Based on the Crop Area Estimating

WANG Ni(Nanjing Forestry University,Nanjing 210037,China)   

  • Online:2011-08-20 Published:2011-08-20

摘要: 以海安县的LANDSAT5—TM影像为研究对象,用于农作物种植面积遥感提取,但由于大气辐射使卫星遥感数据引起畸变从而影响提取精度,为消除该影响,采用黑暗像元减法(DOS)、大气辐射传输模型法FLAASH、ATCOR2、6S等四种大气校正方法进行大气校正并对各种方法进行验证比较,从综合方面考虑得出最优大气校正方法。利用40个主要农作物的样本点的NDVI值与经过大气校正MODIS地表反射率的NDVI值比较,同时利用农作物的反射率值与标准值进行比较对比,以及结合各类主要农作物分类结果精度进行验证。实验结果表明,大气校正在一定程度上提高了农作物真实地表反射率,使得农作物的反射率更加精确,分类精度较原图像也有较大提高,因此提高了主要农作物各类面积提取的精度,综合计算效率和效果等指标得出在面向农作物面积遥感提取的大气方法中,采取FLAASH方法可以达到较好的效果。

Abstract: This paper takes Landsat-TM remote sensing imagery about Haian county as the research object.It corrects the TM remote sensing image by four atmospheric correction methods such as the dark-object subtracting method(DOS) and the methods based on atmospheric radiant transfer models such as FLAASH,ATCOR2,6S.It is hoped to find the best methods of atmospheric correction,which is evaluated from every all-around part such as the efficiency and the precision.First of all,TM is used to operate radiance calibration which transforms a DN value of one pixel to its reflectance value.Secondly,it inputs the reflectance values to FLAASH module for getting the image which composed with the real earth reflectance values.The ATCOR2 and 6S have the same flow.They all need to know the locations of objects,the type and the parameters of sensor,the atmospheric parameters and aerosol values and so on.At the same time,the DOS method only depends on DN values of the TM data.Finally,it is the importance of the research to validate the four methods.And using the classification precisions of the Maximum Likely Classification method,they are compared with TM original classification image,which is the best atmospheric correction method with the highest total precision and Kappa coefficient.Through the atmospheric corrections on a certain extent,it receives the realized earth reflectivity and improves the accuracy of the reflection values about rice in advancing the accuracy of crops estimating area.The result showed that,both ATCOR2 and the FLAASH atmospheric correction are good,and both have the large scale enhancement to the recognition precision about crops.But ATCOR2 has a very long computation process,also the efficiency is not high,in which many experience patterns.The counting efficiency of FLAASH is high,and the reflection value about rice equals to the experimental true value basically.Moreover it has the very big help to improve the precision of recognizing crops area.Therefore the FLAASH atmosphere adjustment method can achieve the best effect regarding the crops area extraction.Summarizing some standards like calculated efficiency and results,the FLAASH atmospheric correction methods win better impacts.

中图分类号: 

  • S127