Circular object recognition from satellite images

Pattern recognition has been the object of interest for many researchers in different fields. In remote sensing, geo-spatial patterns such as circular structures can be observed from satellite imagery and aerial photographs. Recognizing circular structures is difficult in remote sensing imagery. In this article, a detection method of circular forms in satellite image is investigated. Based on Adaboost and Haar features, the method is a machine learning technique. Training data sets are extracted from GoogleEarth. Experiments and results are provided.