A New Two Parameter CFAR Ship Detector in Log-Normal Clutter

Traditional CFAR detectors assume that the statistical model of the sea clutter as a certain distribution, and the models are established through parameter estimation using all the pixel samples in the background window, the modeling precision is influenced by interfering ship targets in the background window. As a result, the parameters will be overestimated, which will cause a degradation of probability of detection (PD), especially in crowded harbors and busy shipping lines. In this paper, a new two parameter CFAR detector in Log-normal clutter is presented. The new two parameter CFAR detector uses log-normal model to fit the gray intensity distribution of the background clutter, by clutter truncation in the background window, the interfering ship targets are removed from the clutter sample, so Log-normal model is precisely built, Compared with traditional CFAR detectors, the parameter estimation is simple and precise, and ship targets in multiple target environment can be also detected. Under the same probability of false alarm (PFA), the proposed two parameter CFAR detector has the highest PD. The superiority of the proposed two parameter CFAR detector is validated on the multi-look Envisat-ASAR data.