Investigation of classification of agricultural areas by using the artificial bee colony algorithm

Turkey is a rich country when it comes to agricultural areas. Therefore it is substantial to gain information about the usage of these areas. Image classification processes with remote sensing has a large share in the extraction of information from satellite images. Classification of agricultural areas from satellite images provides various conveniences in terms of gain information about these areas. It is possible to detect agricultural areas and extract information about which products are grown in which areas through the satellite images. Various methods are used in remote sensing for image classification process. These methods can be sorted as minimum distance, maximum likelihood and parallelepiped method etc. AI optimization algorithms that recently emerged shows the feature of being an alternative to conventional methods. In this thesis, very high resolution satellite images classified using artificial bee colony algorithm. Thus, performed explorations about investigations of classification accuracies of ABC. Necessary control points and training data were collected by using land and aerial images. Rize was chosen for the pilot region as a study area by taking into consideration the vegetation, land topography and agricultural areas. At the end of the application accuracy values that obtained by using the artificial bee colony algorithm were compared with the artificial neural network and maximum likelihood algorithm. It has shown that artificial bee colony algorithm give better accuracy rates than other methods.