Locating the urban area in satellite images to detect changes in them

Satellite images proved to be useful for determining the degree of urbanization. Using them, the growth rate of the urban area can be periodically measured by analyzing the images of the same area. High resolution satellite images provide sufficient detail for this operation. However, their manual handling is very difficult and time consuming. Therefore, an automatic system is necessary. To solve these problems, we propose two novel methods in this study. Our first method is based on urban area detection. Therefore, we first propose an urban area detection method using satellite images. Our urban area detection method is based on the assumption that image contrast will be higher in urban regions compared to other regions. Based on this assumption, our method uses the gradient information of the image as an indicator of the urban area. Then, we use the urban regions detected in bitemporal images to detect changes. Our second method is based on the Histogram of Oriented Gradients (HOG). In this approach, we calculate the histogram of bitemporal images. Then, we define the change detection problem in terms of histogram distance measures. We tested our methods on a diverse data set and provide the results.