Unsupervised segmentation of agricultural regions using TerraSAR-X images

The framework of this study is focused on automatic fast recognition of agricultural interest for TerraSAR-X images. The intended goal is to label regions in an image as fast as possible, into classes significant for a given application, like crop classification. First, a filtering technique is applied to obtain the restored image. Then, two different methods of unsupervised segmentation are used. The Otsu´s method which is based on the optimum threshold of histogram and the k-means method which is based on the Euclidean distance.