A novel signal separation algorithm for a wide band direction-finding system in the HF band

This article describes a novel signal separation algorithm for the purpose of separating the signals overlapped both in time and frequency domain in HF Band by using direction-finding (DF) data. According to the HF wide band DF data, the time-frequency-power spectrum and its corresponding azimuth information are first transformed into azimuth-frequency-energy spectrum of DF data along the time axis; then the open and close operations of morphological image processing are applied to the azimuth-frequency-energy spectrum and isolated noise points are eliminated; finally, the region growing method of the image processing is used Through raw HF wide band DF data validation, this method is proved to be capable of realizing the separation of overlapped signals.