Composite Interference Signals Recognition Based on YOLOv5
Shumin Tu, Qin Cheng, Fei Qian, Pei Hui, Lei Guan · 2022
Complex interference signal identification has been a paramount concern, yet there are very few effective methods. In this work, we propose a composite interference signal identification method based on target detection, which uses a detection box to detect the single interference signal in the composite interference signals (CISs) without separating CISs. This method based on YOLOv5 network can avoid the problem of difficult separation of CISs and be not affected by the multiple characteristics of the composite signal. It can not only effectively identify the composite signal, but also obtain the position information of each single signal. The experimental results validate our analysis. Moreover, it can be concluded that our proposed method is fault-tolerant and easy to migrate.