Automated change detection from high-resolution remote sensing images

Abstract

A fast detection of change in areas of crises or catastrophes is an important condition for planning and coordination of help. This paper describes the results of a cooperative suite of algorithms for automated change detection based on the availability of new satellites with high temporal and/or spatial resolutions. The methods are based on frequency and texture analysis, and segmentation. For the frequency analysis, different band pass filters are applied to identify relevant frequency information for change detection. After transforming the multitemporal images via a fast Fourier transform and applying the most suitable band pass filter, four different methods are available to extract changed structures: differencing and correlation in the frequency domain and correlation and edge detection in the spatial domain. For the texture analysis, we calculate four different parameters (i.e. energy, correlation, contrast and inverse distance moment) for the multitemporal images. The next step is the application of several change detection methods (difference, ratio, regression and principal component analysis) to visualize the changes in the texture images. This method can be combined with a prior segmentation of the image data as well as with morphological operations for a final binary change result. A rule-based combination of the change algorithms is applied to calculate the probability of change for a particular location. The methods were tested with high-resolution satellite images of the crisis areas of Darfour and Haiti. For the frequency based change detection, best results were obtained with adaptive band pass filtering and subsequent edge detection. For the texture based method, a bitemporal principal component analysis for the feature energy provided the best results for change visualization. The next steps will involve the extension of the developed algorithms to test their suitability for other applications such as environmental or phenological change.