Title: Target imaging technology of wireless orbital communication radar

Authors: Xin Tan; Chaoqi Wang; Mingwei Wang; Wenyuan Liu; Xianghui Wang

Addresses: School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xian, Shaanxi, China; Shaanxi Artificial Intelligence Joint Laboratory, Shaanxi University of Science and Technology, Xian, Shaanxi, China ' School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xian, Shaanxi, China; Shaanxi Artificial Intelligence Joint Laboratory, Shaanxi University of Science and Technology, Xian, Shaanxi, China ' School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xian, Shaanxi, China; Shaanxi Artificial Intelligence Joint Laboratory, Shaanxi University of Science and Technology, Xian, Shaanxi, China ' School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xian, Shaanxi, China; Shaanxi Artificial Intelligence Joint Laboratory, Shaanxi University of Science and Technology, Xian, Shaanxi, China ' School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xian, Shaanxi, China; Shaanxi Artificial Intelligence Joint Laboratory, Shaanxi University of Science and Technology, Xian, Shaanxi, China

Abstract: With the continuous advancement of technology, non-communication radar imaging is increasingly used in aerial observation and ground observation fields. This article conducts a systematic study on wireless orbit communication radar imaging technology and compensates for deviations in image quality accuracy through motion autofocus. The results show that the video-based backward projection algorithm proposed in this article is better than the RD algorithm and CBP algorithm in identifying large and small corner points. The feature fusion difference test of the rotated convolution unit was conducted on the ResNet18 and VGG16 neural networks and it was found that the recognition rate of the network architecture using the rotated convolution unit was significantly improved. The designed lightweight network based on rotational convolution units has an average recognition rate of 99.48% in the MSTAR data set.

Keywords: wireless communication radar; radar imaging technology; target recognition imaging; SAR imaging; Bayesian imaging.

DOI: 10.1504/IJGUC.2024.136722

International Journal of Grid and Utility Computing, 2024 Vol.15 No.1, pp.31 - 43

Received: 07 Feb 2023
Accepted: 31 Mar 2023
Published online: 19 Feb 2024 *

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