Title: Analysis of deep learning-based models for automatic target recognition using synthetic aperture radar imagery

Authors: Baldivya Mitra; Maroti Deshmukh; Abhimanyu Kumar

Addresses: National Institute of Technology, Uttarakhand, India ' National Institute of Technology, Uttarakhand, India ' National Institute of Technology, Uttarakhand, India

Abstract: Automatic target recognition (ATR) has become necessary for the military to enhance intelligence in autonomous weapons, ultimately minimise collateral damage, and autonomously tackle war conditions. Therefore, researchers from military weapons manufacturing organisations and academia built various techniques for automatic target recognition models. It is found that synthetic aperture radar (SAR) imagery for automatic target recognition produces a better accuracy of target recognition than other sources of data (Kechagias-Stamatis and Aouf, 2021; Ren et al., 2021). This paper reviews existing techniques for automatic target recognition in autonomous weapons using SAR imagery to analyse their performance in standard operating conditions (SOC) and extended operating conditions (EOC). The experimental results of the state-of-the-art techniques on the MSTAR dataset are compared and analysed. Also, this paper presents the benefits, limitations, and research scopes of existing techniques of ATR.

Keywords: automatic target recognition; ATR; synthetic aperture radar; SAR; moving and stationary target acquisition and recognition; MSTAR.

DOI: 10.1504/IJISTA.2026.152035

International Journal of Intelligent Systems Technologies and Applications, 2026 Vol.24 No.1, pp.54 - 78

Received: 23 Jun 2023
Accepted: 04 Sep 2023
Published online: 04 Mar 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article