Extraction Of Rice Cropping Activities Using Multi-Temporal Ndvi Data And Isodata Clustering-A Case Study Of Raichur District, Karnataka, India.

Abstract

Remote sensing data acquired during a cropping season can assist in assessing crop growth and development performance, and provide information for crop management. Delineating rice cropping is important for crop management and crop production estimation. In this study, we used time series MODIS data to delineate rice cropping activities in Raichur district. The data were processed using the ISODATA clustering, which is a method of unsupervised classification in which algorithm splits and merges clusters. The 250m MODIS NDVI (Normalized Differential Vegetation Index) images of temporal resolution 15 days were stacked and classified in to 100 classes using ISODATA clustering. From the analysis of classified NDVI data, we mapped 1) Crop and Non crop information. 2) Kharif, Rabi and Double crop information. 3) Rice crop information. The total area of Raichur is 8432.8081km2 in which cropland covers area of 6873.89 km2 and remaining 1558.9184 km2 area is covered by non crop. Cropland classified in to, Kharif, Rabi and Double crop which covers an area 1773.4966 km2, 3252.8635 km2 and 1847.5296 km2 respectively. Our study demonstrates potential of multi temporal images that were taken to differentiate and delineate rice cropping activities at a good level of accuracy in spite of the cloudy conditions.
Key words: NDVI; ISODATA; Rice crop; GIS