Identification of Crop Types with the Fuzzy Supervised Classification Using AWiFS and LISS-III Images

Abstract

Proper crop identification is a difficult task in a particular area by using the survey method, and it is essential for planning and management of water resources. Types of crops may be dynamic in any area during a season, and it also depends on the cultivator. The current study is the identification of spatial crops in Rabi season using multi-temporal satellite imagery. Different types of crops have different reflectance in the electromagnetic spectrum. A single crop shows different reflectance in different time of the growing period. The region of interest for the analysis is a part of the Narmada River basin of Madhya Pradesh covering an area of about 20,558 km(2). Land use classification was carried out using LISS-III satellite imagery and demarcation of the agricultural lands of the study area was done. Multi-temporal AWiFS satellite data (9 images) were used for extracting the crop types from October 23, 2011 to March 10, 2012 and NDVI method was applied to the AWiFS satellite data. Different crops have shown different NDVI values during Rabi season. Wheat and soyabean were observed to be the major crops during Rabi season in the study area. Locations (sample points) of different crops were taken from the field survey at the same time period (Rabi). Fuzzy Supervised Classification (FSC) method was applied for identifying the major five crops using NDVI outputs. The results of the accuracy assessment give the overall accuracy and Kappa statistics about 85.60 % and 0.79 respectively.