Change detection in sarimages based on artificial bee colony optimization with fuzzy C - Means clustering

Abstract

Synthetic aperture radar (SAR)generates images with high resolution in all weather conditions for a given application. An Artificial Bee Colony (ABC), optimization algorithm is proposed to detect changes in multitemporal SAR images which are captured at same area in various times. It is well-known fact that the speckle noiseis existed in SAR images.In order to reduce the speckle noise in the co-registered images, a novelAdaptive Median filter is implemented in this paper. Afterthe minimization of speckle noise, discrete wavelet (DWT) fusion is exploited for further image segmentation. Also, an Artificial Bee Colony (ABC) optimization technique is adopted for effective smoothing the image to make decisiveimage classification. Using fuzzy c-means clustering classificationwe can detect changed pixels and unchanged pixels.Finally, theresults are comparedwith DWT-FCM (without optimization), GeneticAlgorithm (GA)optimizationand proposed ABC optimizationAlgorithm. The performance of proposed technique iscomparedin terms ofaccuracy, sensitivity, precision and F1 – score.