Study On CFAR Target Detection Method In Sea Clutter

Target detection in sea clutter background is an important topic in radar signalprocessing. The design and selection of detectors in sea clutter background are rathercomplicated, due to the complicated statistical characteristics of sea clutter. Thedissertation focuses on the performance comparison of the constant false alarm rate(CFAR) detectors in sea clutter background by using real high resolution sea clutterdatasets.At first, the statistical characteristics of sea clutter are introduced. Some models onthe amplitude distributions of sea clutter are reviewed. Moreover, the correlationfunctions and power spectrum density of sea clutter are examined.Secondly, the computer simulation of sea clutter data is the foundation of theperformance analysis of the target detection methods. Two commonly-used algorithmsto simulate the K-distribution clutter, the ZMNL and SIRP, are realized by the computersoftware. The simulated results show that the software can perform well the simulationof the K-distributed clutter data.Thirdly, the non-coherent and coherent constant false alarm rate (CFAR) detectionmethods are discussed. Under the Gaussian clutter assumption, the CFAR thresholds ofthe CA-CFAR、COA-CFAR and CM-CFAR detectors are derived theoretically, whereCOA-CFAR detector and CM-CFAR detector are improved versions of the CA-CFARdetector. Due to the non-Gaussian nature of real sea clutter and complicated targetenvironment (for example, detection problem in dense target regions), the performanceof the above-mentioned three detectors must be re-evaluated. By using the IPIX radarsea clutter datasets, the detection performance of the CA-CFAR、COA-CFAR and CM-CFAR detectors are compared. The experimental results show that the CM-CFARdetector can always attain better target detection performance whether non-coherentintegration or the coherent integration when there exist neighboring strong interferencefrom other targets. These conclusions are important for selecting the detection methodsin marine radars.