Unsupervised classification and analysis of Istanbul-Turkey satellite image utilizing the remote sensing

• Main Authors: Mohammed, Ali Abdulwahhab, Al-Ghrairi, Assad H. Thary, Al-zubidi, Azhar F., Saeed, Harith M.
• Format: Conference Proceeding
• Language: English
• Subjects: Agriculture; Buildings; Classification; Image classification; Remote sensing; Satellite imagery; Spatial resolution
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0118339

The present study had used Landsat5 satellite images, which is a polar-orbiting, multi-spectral high 30m spatial resolution for covering Istanbul with 3 combination bands, which are: Blue (Band1), Red (Band3), and Green (Band2). This study has the aim of the classification and analyses of the Istanbul-Turkey study area for 2004 and discussing the methods which are utilized for the computation of the area cover of every one of the classes. The present study included 2 phases, which are: training and classification. In the 1st stage, features are obtained with the use of the moment based k-means approach in the Remote Sensing (RS) and stored it in the data-set. The 2nd stage using the extraction of the moments with the K-Means for image classification, in which, results of classification exhibit 5 areas (which are: Agriculture or forest, sea, buildings with the Agriculture, bare lands and buildings with no Agriculture) where, it may be noticed that the area has been secured by every one of the classes. Experimental results of the method of the classification showed sufficient performance precision with a good classification and result analyses of area that has been covered for every one of the classes or regions.