Investigation of the effects of different image classification and enhancement methods on land use/land cover mapping

With Remote Sensing technologies, land use/land cover information can be obtained in a very short time, at lower costs and with more objective measures than classical methods. Remote sensing data is an important component of LU / LC studies. The general aim of the study is to analyze and model the current state of the LU / LC of Samandağ district of Hatay province, which is the study area, using remotely sensed images and different classification techniques. To achieve the main goal of the study: (a) Use of different classification techniques ( supervised, unsupervised, and object-based classification) on high resolution satellite images and comparison and analysis of classification results; (b) Use of different image enhancement and enrichment techniques and analysis of effects on accuracy analysis results, and (c) Modeling, monitoring and mapping LU / LC of study area. Using the principal component analysis and minimum noise partition image enhancement techniques in the study, the effects of these applications on the accuracy analysis were investigated. According to results obtained in this study, the object-based classification method gave the highest accuracy with 80.45% and 0.74 kappa statistic.