Vol.:(0123456789) 1 3 Journal of Ambient Intelligence and Humanized Computing https://doi.org/10.1007/s12652-020-02091-y ORIGINAL RESEARCH A novel synthetic aperture radar image change detection system using radial basis function‑based deep convolutional neural network B. Pandeeswari 1  · J. Sutha 2  · M. Parvathy 1 Received: 28 February 2020 / Accepted: 5 May 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020 Abstract Today, the automatic change detection and also classification as of the Synthetic Aperture Radar (SAR) images remain a hard process. In the existing research, the availability of Speckle Noise (SN), high time-consumption, and low accuracy are the chief issues. To resolve such issues, this paper proposed a novel SAR image change detection system utilizing a Radial Basis Function-based Deep Convolutional Neural Network (RBF-DCNN). The proposed methodology comprises six phases, namely, pre-processing, obtaining difference image, pixel-level image fusion, Feature Extraction (FE), Feature Selection (FS), and also change detection (CD) utilizing the classifier. Initially, the noise is eliminated as of the input, SAR image 1 and SAR image 2, utilizing the NLMSTAF approach. Subsequently, the difference image is attained by utilizing a Log-ratio operator (LRO) and Gauss-LRO, and the attained difference image is then fused. Next, the LTrP, WST, edge, and MSER features are extracted from the fused image. As of those features that were extracted, the necessary features are selected utilizing the Hybrid GWO-GA algorithm. The features (selected) are finally inputted to the RBF-DCNN classifier for detecting the changes in an image. Experimental outcomes established that the proposed work renders better performance on considering the existing system. Keywords Non local mean spatio temporal adaptive filtering (NLMSTAF) · Log-ratio · Gauss-log-ratio operator · Local tetra pattern (LTrP) · Wavelet statistical transform (WST) · Maximally stable external region (MSER) · Hybrid Gray Wolf optimization-genetic algorithm (Hybrid GWO-GA) · Radial basis function-deep convolutional neural network (RBF- DCNN) 1 Introduction The process of recognizing the changes that occur between the two imageries of the same scene but taken at disparate times is termed as Image CD (Li et al. 2015). It is a signifi- cant issue in the military and civil fields (Gong et al. 2016). Multi-temporal RS imagery performs a crucial role in copi- ous fields of applications (Yousif and Ban 2016). But the contemporary CD system utilizes the SAR imageries (Jia et al. 2015). The SAR stands as active microwave coher- ent imaging radar; hence, it can obtain the Remote Sensing (RS) data every day, irrespective of the weather. Further- more, it could make-up on behalf of the lack of optics and infrared RS (Hou et al. 2014). Its application includes envi- ronmental monitoring, urban study, disaster management, et cetera. SAR is insensitive to atmospheric as well as sun illumination conditions; consequently, SAR is appropriate for the CD tasks (Jia et al. 2014). On comparing with con- ventional optical RS images, it is clear that a SAR image can be effectively captured under all illuminations and weather conditions (Zheng et al. 2014). Research has been going over SAR image CD techniques for the past decades in the unsupervised, supervised, and semi-supervised manners. As unsupervised approaches do not need the prior information of the source images, they have grabbed researchers’ atten- tion towards them (Wang et al. 2016). * B. Pandeeswari pandeeswari.sit@gmail.com J. Sutha sutha_skad@gmail.com M. Parvathy Parvathydurai2002@gmail.com 1 Department of Computer Science and Engineering, Sethu Institute of Technology, Kariapatti, Virudhunagar, India 2 Department of Computer Science and Engineering, AAA College of Engineering and Technology, Sivakasi, India