Translator Disclaimer
12 December 2006 Paddy ground truth data collection and evaluation for land cover mapping
Author Affiliations +
Ground truth data have grid cells that consist of single land cover classification (pure grid cell), and they are useful as the training data in classifying global land cover. Previous studies showed many global land cover map, however, they have rarely reported locations and collection method of the training data that they might use. Paddy is indispensable classification for global land cover mapping but there is no established method for collecting paddy ground truth data. In this study we collected paddy ground truth candidate data from the existing 1km-resolution China national land use data, which was produced using 30m-resolution LANDSAT/TM data. Since the land use data recorded classification area ratio in each grid cell, it is efficient to collect paddy's pure grid cells. We collected pure grid cells of paddy and segmented them into small areas, and obtained paddy ground truth candidate sites. After the sites whose area >= nine grid cells (=9 km2) were selected, we examined the selected sites' Normalized Difference Vegetation Index (NDVI) time-series changes in 2003 using TERRA/Moderate Resolution Imaging Spectroradiometer (MODIS) 16-days composite data. Furthermore, we chose the sites whose NDVI standard deviation =< 0.1 through 2003, as a result, the number of the selected sites was 271. The 271 candidate sites were assigned to nine China climatic zones, and the sites that are located near climatic boundaries were eliminated. The site whose area is largest in each climatic zone was tried to be selected. Finally, ten sites of paddy ground truth data were collected.
© (2006) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Hiroshi P. Sato, Ryutaro Tateishi, and Jieying Xiao "Paddy ground truth data collection and evaluation for land cover mapping", Proc. SPIE 6411, Agriculture and Hydrology Applications of Remote Sensing, 641120 (12 December 2006);

Back to Top