An Improved 3-D Reconstruction Method based on Deep Neural Network
Conference: EUSAR 2022 - 14th European Conference on Synthetic Aperture Radar
07/25/2022 - 07/27/2022 at Leipzig, Germany
Proceedings: EUSAR 2022
Pages: 5Language: englishTyp: PDF
Authors:
Zhang, Tianyi; Liu, Siyuan; Wei, Yangkai; Wang, Guanxing; Gao, Yongpeng (Radar Research Lab., School of Information and Electronics, Beijing Institute of Technology, Beijing, China &b Beijing Key Laboratory of Embedded Real-time Information Processing Technology, Beijing Institute of Technology, Beijing, China)
Ding, Zegang (Radar Research Lab., School of Information and Electronics, Beijing Institute of Technology, Beijing, China &b Beijing Key Laboratory of Embedded Real-time Information Processing Technology, Beijing Institute of Technology, Beijing, China & Beijing Institute of Technology Chongqing Innovation Center, Chongqing, China)
Zhang, Yan; Guo, Xin (DFH Satellite Co. Ltd, Beijing, China)
Abstract:
Current three-dimensional (3-D) reconstruction methods based on two-dimensional (2-D) inverse synthetic aperture (ISAR) image sequences usually consist of some sequential nonlinear steps, and they face with the error accumulation and transmission inevitably. To realize precise target reconstruction, an improved 3-D reconstruction method based on motion parameters and deep neural network (DNN) is proposed. The proposed method could realize the end-to-end transformation from the motion parameters to the 3-D target via DNN, and the error transmission and accumulation can be avoided. Results based on the synthetized data set validate the proposed method.