一种基于可学习权重的低频雷达超分辨成像方法和装置
By constructing a band interpolation network with learnable weights and using a supervised learning method, the problem of low band interpolation accuracy in low-frequency imaging radar was solved, and high-resolution radar image generation was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TSINGHUA UNIVERSITY
- Filing Date
- 2022-12-20
- Publication Date
- 2026-07-17
AI Technical Summary
Existing low-frequency imaging radars suffer from low frequency interpolation accuracy and consequently low radar imaging resolution because their frequency interpolation algorithms do not consider different frequency dependence factors.
A deep learning-based frequency band interpolation network is adopted. By constructing a frequency band interpolation network with learnable weights, the optimal learnable weights are trained using supervised learning methods. The interpolation is then performed in combination with an improved fast iterative thresholding algorithm to balance the influence of different frequency dependence factors on frequency band interpolation.
It improved the accuracy of frequency band interpolation, enhanced radar imaging resolution, and enabled the generation of high-resolution radar images.
Smart Images
Figure CN116184399B_ABST