A two-stage classification algorithm for radar targets based on compressive detection

Abstract Algorithms are proposed to address the radar target detection problem of compressed sensing (CS) under conditions a low signal-to-noise ratio (SNR) and signal-to-clutter (SCR) echo signal. The algorithms include two-stage classification for targets based on compressive (CD) without signal reconstruction support vector data description (SVDD) one-class classifier. First, we present sparsity in distance dimension design measurement matrix CD Constant false alarm rate (CFAR) is performed directly complete first-order classification. In simulations, performance similar that traditional matched filtering algorithm, but lower, necessary storage space reduced. Then, power spectrum features extracted from after converted feature domain. SVDD classifier introduced train classify characteristic signals separation alarms. Finally, algorithm verified by simulation. number alarms reduced, probability improved.