Examination of remote sensing methods in change detection analysis. case of study: Bodrum peninsula

This doctoral thesis emphasizes the critical importance of monitoring land use changes in urban areas over time, focusing on the utilization of remote sensing (RS) data and image classification methods. The research is conducted to examine the changes in land use over a period of 30 years in the touristic Bodrum Peninsula, Turkey. In the initial stage of the thesis, a literature review and evaluation of 10 different image classification methods were carried out. Maximum Likelihood Classification (MLC), Support Vector Machines (SVM), and Artificial Neural Network (ANN) methods were employed to classify five classes (forest, agricultural land, settlement, water, and bare lands) for the analysis of changes in the study area. The classification results using satellite images revealed that MLC achieved high accuracy in forest classification. SVM emerged as an effective method for agricultural areas, providing acceptable results for settlement and bare lands. ANN demonstrated an advantage in detecting finer details compared to other methods. The findings indicate a threefold increase in settlement areas over 30 years in the Bodrum Peninsula, with a decrease in forests and agricultural areas in the first 20 years and increas in the last 10 years. Different trends were observed for the water class using various methods, while a general increase in open spaces persisted until 2010. By evaluating of MLC, SVM, and ANN methods in urban change, this study provides valuable information for environmental, and urban management. It is anticipated that this research will serve as a fundamental reference for future studies.