ON THE PERFORMANCE OF SUPERVISED CLASSIFIERS FOR CROP IDENTIFICATION AND ESTIMATION USING MULTI-SPECTRAL IMAGERY
Asif Ali Khan*, Nasru Minallah*, Shahbaz Khan**
ABSTRACT
The objective of this research is to investigate crop estimation using SPOT-5 satellite imagery. We specifically considered
tobacco as our pilot crop and compared the obtained results with manually delineated calculations. For
this research SPOT-5 imagery of 2.5m spatial resolution, was provided by Space and Upper Atmosphere Research
Commission (SUPARCO), space agency of Pakistan. After preprocessing, which is a preparatory step in analyzing
and classifying satellite imagery to improve classification results and reduce the efforts and processing time, different
supervised classifiers namely Maximum Likelihood approach, Neural Network and Minimum Distance Classifier have
been used to classify the imagery. Training data for classifiers has been collected through multiple field surveys using
GPS receivers. The results obtained clearly show that the performance of maximum-likelihood classifier is better than
the other considered counterparts. Also it is indicated that the newly developed system offer an efficient, reliable and
faster approach for estimation of tobacco crop.
To share on other social networks, click on any share button. What are these?
June 2015
Vol 34 No 1 (2015)
