基于图像分割的摩尔斯信号检测算法

Morse Signal Detection Algorithm Based on Image Segmentation

摩尔斯信号在通信领域应用广泛,具有十分重要的研究价值。但是,由于短波信道比较复杂,环境恶劣,信号质量差,短波宽带条件下的摩尔斯信号检测问题一直没有较好的解决方案。基于此,提出了一种基于时频图和图像分割的摩尔斯信号检测算法。算法主要分为粗检测和识别两部分,粗检测是指基于图像分割的方法,从宽带数据中初步检测出摩尔斯信号;识别部分是指通过提取最大持续长度等特征,逐级排除非摩尔斯信号,进一步识别摩尔斯信号。实际信号测试表明,该算法对摩尔斯信号的正确检测率为96%以上,虚警率小于7.2%,相较于对比算法整体性能较好,具有一定的工程应用价值。

Morse signal is widely used in the field of communication and has very important research value.However,due to the complexity of HF channel,bad environment and poor signal quality,there is no good solution to the problem of Morse signal detection in HF Broadband.Based on this,this paper proposes a Morse signal detection algorithm based on spectrogram and image segmentation.The algorithm is mainly divided into two parts:rough detection and recognition.Rough detection is a method based on image segmentation to initially detect Morse signals from broadband data;recognition part is to extract the maximum duration and other features to exclude non Morse signals step by step and further recognize Morse signals.The actual signal test shows that the correct detection rate of Morse signal is more than 96%,and the false alarm rate is less than 7.2%.Compared with the contrast algorithm,the overall performance of the algorithm is better,which has a certain engineering application value.