Research on The Application of Attention U-Net to Image Sorting of Radar Signals
In order to face the challenge of radar signal sorting in the complex electronic environ-ment of naval battlefield,this paper proposes a new sorting method by using an improved U-Net network combined with attention mechanism.Firstly,pulse description words are transformed into image sequences to adapt to deep learning processing.By optimizing U-Net architecture and in-tegrating attention mechanism,the model can effectively improve the recognition & extraction abil-ity of key pulse features and realize pixel-level classification.In this way,the system can accurately search and classify all radar pulses.Experimental results show that the proposed method signifi-cantly improves the accuracy of radar signal sorting in the complex electromagnetic environment of naval battle field,and provides an efficient solution to deal with strong interference environment.This research result confirms the superiority and practicability of Attention U-Net in radar signal intelligent sorting.
radar signal sortingU-Net networkattention mechanismpulse description word
NETL
NSTL
万方数据
