Target Detection based on Two-dimensional Fractal Property under Sea Clutter Background

To overcome the shortages of the traditional target detection method based on a statistical model, this paper concentrate on the fractal property of sea clutter and its application on target detection. Since the complex properties of sea clutter, it is hard to detect targets based on a single fractal parameter. Therefore, this paper analyzed the two-dimensional fractal properties of sea clutter, where the box-counting dimension and fractal model fit error of autoregressive (AR) spectrum are treated as the feature inputs for target detector. Then, real measured S-band sea clutter datasets are taken to analyze the two-dimensional property of sea clutter and several datasets are taken to test the performance of the detection method. Finally, from the analysis of real sea clutter datasets, the proposed method based on two-dimensional fractal property has a better detection performance than the traditional CFAR method and existing fractal methods.