Study of Land Cover Changes of Baghdad Using Multi-Temporal Landsat Satellite Images

Abstract

The main goal of this work is study the land cover changes for "Baghdad city" over a period of (30) years using multi-temporal Landsat satellite images (TM, ETM+ and OLI) acquired in 1984, 2000, and 2015 respectively. In this work, The principal components analysis transform has been utilized as multi operators, (i.e. enhancement, compressor, and temporal change detector). Since most of the image band's information are presented in the first PCs image. Then, the PC1 image for all three years is partitioned into variable sized blocks using quad tree technique. Several different methods of classification have been used to classify Landsat satellite images; these are, proposed method singular value decomposition (SVD) using Visual Basic software and supervised method (Maximum likelihood Classifier) using ENVI 5.1 software are utilized in order to get the most accurate results and then compare the results of each method and calculate the land cover changes that have been taken place in years 2000 and 2015; comparing with 1984. The image classification of the study area resulted into five land cover types: Water body, vegetation, open land (Barren land), urban area "Residential I" and urban area "Residential II". The results from classification process indicated that water body, vegetation, open land and the urban area "Residential I" are increased, while the second type from urban area "Residential II" in decrease to year 2015 comparable with 1984. Despite use of many methods of classification, results of the proposed method proved its efficiency, where the classification accuracies for the (SVD) method are 81%, 78% and 80% for years 1984, 2000 and 2015 respectively.