Downloads: 136 | Views: 242

Research Paper | Astronomy Science | Iraq | Volume 7 Issue 6, June 2018

Classification of Al-Chabaish Marshes Satellite Images in Iraq using Support Vector Machine Technique

Saad M. Hel | Bushra Q. Al-Abudi ^[2]

Abstract: In this paper, Landsat satellite images of AL-Chabaish marshes and surrounding district in (Dhi Qar) province in the south of Iraq are classified for years 1990, 2000, 2015 and 2017 using two software programming these are MATLAB r2017a 2017 and ENVI 5.3 2017. Principle components analysis (PCA) is applied on six bands of these satellite images using MATLAB and the information of all six bands concentrated in first three principle components and then blended to form integrated image. Then the integrated image is partitioned using Adaptive Horizontal-Vertical partitioning scheme, finally supervised Support Vectors Machine (SVM) technique is applied to classify the image. While supervised method (Maximum likelihood) using ENVI software is utilizing to compare the results. The results from using SVM provided high accuracy these are, 90.98 %, 93.21 %, 92.17 % and 91.13 % for years 1990, 2000, 2015 and 2017 respectively. The results indicated that the AL-Chabaish marshes from 1990 to 2017 is suffering many artificially and naturally change, this clearly impact in the Landover. In 1990 the marshes drying and this causes reduce in a vegetation, water and increase in openland, shallow water, the effect of drying it was very clear in 2000, 2015, and 2017 the AL-Chabaish landcover situation improved much as the amount of water and vegetation increased with a decreasing in the openland

Keywords: Support Vectors Machine, Landsat satellite images, Principle components analysis, AL-Chabaish Marshes

Edition: Volume 7 Issue 6, June 2018,

Pages: 688 - 695