摘要
This paper mainly revolves the time-frequency image of low probability of intercept(LPI) radar signals and carries out research work on image features selection and extraction and recognition. Since Choi-Williams distribution(CWD) uses the exponential kernel of bilinear generalized class of time-frequency distribution, it has an excellent time-frequency aggregation. And it is suitable for detecting LPI radar signals in a low signal-to-noise ratio(SNR) condition. A radial integration method based on the integral rotating factor is proposed to detect LPI radar signals when the signals' time-frequency image is obtained. First, the digital image processing method is used to preprocess the LPI radar signals' time-frequency images after CWD transformation; then, the radial integration method based on the integral rotating factor is used to detect LPI radar signals in the binary images. The analytic results of real data show that the method has a good performance on detecting LPI radar signals in a low SNR condition. Additionally,the method is simple and takes less logic resources and has the potential of real-time detection of LPI radar signals.
This paper mainly revolves the time-frequency image of low probability of intercept （LPI） radar signals and carries out research work on image features selection and extraction and recognition. Since Choi-Williams distribution （CWD） uses the ex- ponential kernel of bilinear generalized class of time-frequency distribution, it has an excellent time-frequency aggregation. And it is suitable for detecting LPI radar signals in a low signal-to-noise ratio （SNR） condition. A radial integration method based on the in- tegral rotating factor is proposed to detect LPI radar signals when the signals＇ time-frequency image is obtained. First, the digital image processing method is used to preprocess the LPI radar sig- nals＇ time-frequency images after CWD transformation; then, the radial integration method based on the integral rotating factor is used to detect LPI radar signals in the binary images. The analytic results of real data show that the method has a good performance on detecting LPI radar signals in a low SNR condition. Additionally, the method is simple and takes less logic resources and has the potential of real-time detection of LPI radar signals.
