2009, 31(12): 2881-2885.
doi: 10.3724/SP.J.1146.2008.01707
Abstract:
In all the algorithms of the ship detection of SAR images, two parameter CFAR detector uses three moving windows: target window, protect window and background window, the sizes of the three windows and the moving step need to be trained, So it is quite inefficient and will cause targets undetected when they are too close. The improved two parameter CFAR detector delivered in this paper only uses a target window and a background window, By using special methods to remove the leaked ship pixels in the background window and estimate the remaining pixels in the background window which are sea clutter to get the local windows clutter gray mean and variance. The moving step is the same as the length of the target window, Compared with two parameter CFAR detector, the structure is simplified and the detection results false alarms is less, targets too close can also be detected, furthermore, the computing efficiency is improved. The simulation results prove the algorithms effectiveness.