This paper proposes a system for finding changes in satellite images taken at two different intervals of time and it may be some years or occurrences of images before and after flooding etc.,. The changes are obtained as binary labels (changed/unchanged) for each pixel. The input images may be of monochromatic or color images and also these images may be of various graphic file formats like PNG, BMP, JPG etc.,. The RGB images are converted to gray images. The mean ratio and log ratio are obtained from both input images. These are fused together to form a difference image using Discrete Wavelet Transform (DWT). A fuzzy clustering approach is applied to classify both changed and unchanged regions in the difference image. Previously this system is done using Fuzzy Local Information C means clustering. To enhance the changed information and to reduce the effect of noise, the spatial contextual information is added along with the self data information using Reformulated Fuzzy Local Information C-means Clustering (RFLICM). The resultant change detected images are showing good performance than the existing techniques. The performance measures are compared using Percentage of Correct Classification (PCC) and Kappa statistic parameter which gives the measure of accuracy for FLICM and RFLICM. Experimental results are compared for different dimensions of satellite images with different file formats taken at different intervals of time and the computation time is also found out. © 2015 Elsevier B.V., All rights reserved.