395 Real-time Adaptive Approach for Hidden Targets Shape Identifcation using through Wall Imaging System Sandeep Kaushal # , Bambam Kumar @ , Prabhat Sharma $ , and Dharmendra Singh * # Department of Electronics and Communication Engineering, ACET, Amritsar - 143 109, India @ Department of Electronics and Communication Engineering, National Institute of Technology, Patna - 800 005, India $ Instrument Research and Development Establishment, Dehradun - 248 008, India * Department of Electronics and Communication Engineering, Indian Institute of Technology, Roorkee - 247 667, India * E-mail: dharmfec@iitr.ernet.in AbStrAct In Through-wall Imaging (TWI) system, shape-based identifcation of the hidden target behind the wall made of any dielectric material like brick, cement, concrete, dry plywood, plastic and Tefon, etc. is one of the most challenging tasks. However, it is very important to understand that the performance of TWI systems is limited by the presence of clutter due to the wall and also transmitted frequency range. Therefore, the quality of obtained image is blurred and very difcult to identify the shape of targets. In the present paper, a shape-based image identifcation technique with the help of a neural network and curve-ftting approach is proposed to overcome the limitation of existing techniques. A real time experimental analysis of TWI has been carried out using the TWI radar system to collect and process the data, with and without targets. The collected data is trained by a neural network for shape identifcation of targets behind the wall in any orientation and then threshold by a curve-ftting method for smoothing the background. The neural network has been used to train the noisy data i.e. raw data and noise free data i.e. pre-processed data. The shape of hidden targets is identifed by using the curve ftting method with the help of trained neural network data and real time data. The results obtained by the developed technique are promising for target identifcation at any orientation. Keywords: TWI; Neural network; Threshold; Curve-ftting Defence Science Journal, Vol. 71, No. 3, May 2021, pp. 395-402, DOI : 10.14429/dsj.71.16696 © 2021, DESIDOC 1. IntroductIon Through-wall imaging detection and identifcation are one of the most challenging tasks for military applications, surveillance and security of the public and their assets because hidden targets cannot be sensed through conventional techniques. In this domain of these applications, a key challenge for these applications is not only to fnd the presence of various targets behind the wall but also to identify their shape and size. Researchers have implemented ultra-wideband microwave imaging radar, based on the stepped frequency mode, to penetrate through concrete wall materials and make smart decisions about the contents of buildings, but still trying to develop some intelligence techniques to identify the hidden objects. Nevertheless, the deep machine learning and soft computing methods have been used to classify the targets in diferent domain applications 1,2 . As a microwave imaging radar based on the stepped frequency continuous wave (SFCW) technique is commonly applied to various practical applications including civil engineering 3,4 , detection of pipes and cables buried in the ground 5,6 , through-wall radar imaging 7–11 , medical imaging 12 , feature estimation of the road surface layer 13 , and many applications of ground penetrating radar (GPR) 14-17 . SFCW radar also has various advantages such as, high receiver sensitivity and high mean transmitter power. It does not only ofer good capability of detecting the targets but also improves range accuracy, facilitates clutter rejection, and helps in suppressing the multipath propagation 7 . In recent years, many artifcial intelligence problems have been solved using convolutional neural networks (CNNs). CNN ofers a lot of advantages over the other machine learning methods in resolving complicated learning tasks. In conventional identifcation methodologies, various features such as colours, edges and corners are physically extracted and designed and after that support vector machine (SVM) techniques are used as traditional classifer 18, 19 . In literature, through-wall imaging system has been classifed into two categories: one is based on image detection and the other is on feature extraction. The frst category uses the beam focusing algorithms such as back projection, delay and sum, etc. The beam focusing parameter, an accurate wall characteristic estimation, is one of the main challenges of through-the-wall imaging (TWI) system. The propagation velocity of electromagnetic wave (EM) waves is afected by dielectric constant and thickness of the wall. Therefore, the wall parameters such as thickness and dielectric constant needs to be estimated precisely. A small error in the wall parameters Received : 12 January 2021, Revised : 02 February 2021 Accepted : 12 February 2021, Online published : 17 May 2021