Improving the Accuracy of Land Cover Classification using Sentinel 2 Data and Knowledge Based Classification System in the West of Amara City, Iraq

Abstract

In this paper, we focus on the use of remote sensing analysis technique with the Sentinel-2 data in order to discriminate land cover in Al-Maimouna District, West of Amara, Iraq. The main objective of the study is to explore whether the expert classification technique can improve the accuracy of land cover classification. In our experiment, an expert classification technique was applied to the Sentinel-2 data set by using, normalized difference vegetation index, soil-adjusted vegetation index, normalized difference water index, modified normalized difference water index, urban index, normalized difference built-up index, normalized difference soil index and bare soil index. Build-up area was found to be the dominant type of land use classified which covers about 41.9% of the total study area, while the least classified was vegetation which accounts for 15.1%. The results of the study demonstrates that the expert classification technique produced an overall accuracy of 88.3%. The results of the study also indicate that by using an expert classification technique, a significantly higher discrimination accuracy can be achieved. Moreover, the expert classification technique reduces problems associated with high-resolution images; while reliabilities are achieved that are better than those achieved with traditional methods.