Abstract：A modified ACCA-CFAR method with local adaptive threshold is proposed for SAR image detection based on ACCA-CFAR in this paper, in which the K-distribution is used to describe the clutter. First the parameter of K-distribution is estimated, secondly, the local ODV (Ordered Data Variability) threshold is determined and the background pixels are selected, finally the detection is made. The paper also proves that the ODV threshold is related with the parameter of the local statistical model. Finally, the algorithm is applied to SAR image of ship, and the result indicates that there are fewer false alarms in the new algorithm, and the whole ship targets are detected with more structure details.
2009, Vol. 31 
