Staged BOC signal detection method based on cyclostationary characteristics

The invention relates to a signal detection method, in particular to a staged BOC signal detection method based on cyclostationary characteristics. An improved cyclostationary feature detection methodis provided, cyclic spectral density maps of signals and noise are mapped into two-dimensional grey-scale maps respectively, then the two types of images are put into a convolutional neural network (CNN) for self-training to extract features, and then the trained network is utilized to classify the images so as to realize detection. In order to detect signals more efficiently, staged detection isadopted, when the signal-to-noise ratio is large, the signals can be detected rapidly and accurately through energy detection, and the detection time is shortened; and when the signal-to-noise ratiois small, the energy detection is inaccurate, and the cyclic stationary characteristic detection is improved in the second stage, so that the detection probability is improved. 本发明涉及信号检测方法，具体的是一种基于循环平稳特征的分阶段BOC信号检测方法。提出改进循环平稳特征