An Interference Detection Method for Satellite Communications based on Maximum Between-Class Variance of Spectral Histogram

Interference detection is one of the critical prerequisites for avoiding and suppressing interference in satellite communication applications. In this article, we propose to treat the frequency-domain interference detection as a threshold segmentation problem and develop an interference detection method from a novel perspective. First, to deeply analyze the characteristic differences between interference and background noise from a statistical perspective, a Spectral Histogram of mixed signals is introduced. Then, a frequency-domain preprocessing method, which combines a morphological filter and a mean filter, is designed to reduce the impact of background noise. Inspired by the Otsu adaptive thresholding method in the field of image processing, an adaptive interference detection method based on a Maximum Between-Class Variance (MBCV) criterion is finally proposed to detect the interference in a wideband spectrum environment. Simulation experiments have been conducted on interference signals with different types and parameters, and the detection results validate the effectiveness of the proposed method.