ABSTRACT

Field studies were conducted in 2002 at Long View Farm located in east-central Mississippi. Two cotton fields, Field-104 and Shop Field, were selected for the remote sensing of the biophysical properties of the cotton crop. At each field plots were laid out using previous crop normalized difference vegetation index (NDVI) into low, medium and high categories. The objectives of the study were to determine the potential of spectral analysis of multispectral images and field spectroradiometer data for discriminating morningglory and grasses from cotton canopy. Multispectral images were used to derive NDVI tem-poral pattern analysis to discriminate weeds from cotton canopy and feature extraction technique were used to identify and map morningglory on field basis. Images acquired on July 18, and August 28 show lower NDVI values in the field where weeds were intermixed with cotton plants compared to weed-free cotton canopy. Near infrared band (850 nm) of the multis-pectral imagery and spectroradiometer ~ 750 - 1000 nm wavelength region showed promising results for discriminating be-tween weed-free cotton canopy and cotton canopy with morningglory. Spectral extraction process identified and mapped < 50 % of the weeds present in the study area.