A semantic segmentation-based SAR image radio frequency interference suppression method

By employing a semantic segmentation-based SAR image radio frequency interference suppression method, utilizing the UNet network and enhanced attention module, the shortcomings of traditional and intelligent methods in suppressing weak interference in complex electromagnetic environments are addressed. This method achieves high-precision and rapid interference suppression and preservation of useful signals, thereby improving the quality of SAR images.

CN117192490BActive Publication Date: 2026-06-19NORTHWESTERN POLYTECHNICAL UNIV
3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NORTHWESTERN POLYTECHNICAL UNIV
Filing Date
2023-08-26
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing radio frequency interference suppression methods are difficult to effectively suppress weak interference and retain useful signals in complex electromagnetic environments. Traditional methods suffer from missed detections and signal loss, while intelligent methods are prone to false alarms in large areas of negative sample data.

⚗Method used

A semantic segmentation-based SAR image radio frequency interference suppression method is adopted. The UNet network structure and enhanced attention module are used to train the SAR image and generate interference masks through the semantic segmentation network model, and interference suppression is achieved by combining frequency domain notch filtering technology.

🎯Benefits of technology

It improves the accuracy and speed of interference suppression, effectively preserves useful signals, and enhances the quality of SAR images, especially under weak interference conditions.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention relates to a SAR image radio frequency interference suppression method based on semantic segmentation, belonging to the field of radar signal processing. First, the received SAR data is subjected to FFT to obtain a two-dimensional spectrum. The two-dimensional spectrum is then resampled, and the resampled spectrum sample is input into a trained semantic segmentation network model to obtain the interference mask predicted by the network. Based on the final mask, frequency domain notch filtering is applied to the original two-dimensional spectrum, setting the positions where the interference mask is 1 to zero in the spectrum. The interference-suppressed two-dimensional spectrum is windowed according to the required parameters, and then transformed into the time domain through inverse Fourier transform to obtain the final interference-suppressed SAR image. This invention can achieve higher accuracy and faster computation in SAR image interference suppression, and effectively preserves useful signals in the detection and suppression of weak energy interference, significantly improving the quality of the interference-suppressed SAR image.
Need to check novelty before this filing date? Find Prior Art