A study on impacts of residential area in MODIS mixed pixels on corn LAI and FPAR estimation by remote sensing
View Fulltext  View/Add Comment  Download reader
  
DOI:10.7606/j.issn.1000-7601.2010.02.49
Key Words: MODIS  mixed pixel  residential area  LAI  FPAR
Author NameAffiliation
YANG Fei Northeast Institute of Geography and Agroecology of Chinese Academy of Sciences, Changchun, Jilin 130012, China
Graduate University of Chinese Academy of Sciences, Beijing 100039, China 
ZHANG Bai Northeast Institute of Geography and Agroecology of Chinese Academy of Sciences, Changchun, Jilin 130012, China 
LIU Dianwei Northeast Institute of Geography and Agroecology of Chinese Academy of Sciences, Changchun, Jilin 130012, China 
LIU Zhiming College of Urban and Environmental Sciences, Northeast Normal University, Changchun, Jilin 130024, China 
SONG Kaishan Northeast Institute of Geography and Agroecology of Chinese Academy of Sciences, Changchun, Jilin 130012, China 
WANG Zongming Northeast Institute of Geography and Agroecology of Chinese Academy of Sciences, Changchun, Jilin 130012, China 
Hits: 184
Download times: 123
Abstract:
      MODIS data become more and more important source for research and production, but the low spatial resolution leads to the data influenced by many factors. Plenty of works have been done for data precision validation, which demonstrated that mixed pixel of several objects was the main influencing factor. Equal dispersion of the rural residential area is the typical characteristic of the farmland in northeastern China. However, it has not been seen the influences of residential area on the MODIS data precision of mixed pixels. In this study, it was studied the residential area in typical corn field impacts on the precision of MODIS LAI and FPAR data based on TM data. It can be concluded that more than 96% of the MODIS mixed pixels have the residential area below 40%. Residential area made strong impact on the precision of MODIS LAI and FPAR data. The residential area in a MODIS pixel showed great linear relations with the difference before and after considering the residential area for red band reflectance, LAI and FPAR, R2 of which are above 0.90, but with R2 of 0.3423 for near infrared band reflectance, this demonstrated that near infrared band reflectance of MODIS mixed pixels are influenced greatly by residential area.