Warship detection in smoke screen interference based on fractal and rough set theory

In IR images, the same kind of natural scenes are self-similarity statistically, namely their fractal dimension remains unchanged approximately. At edges, however, the dimension is underestimated, because edges add a deterministic component to the surface which is not assumed to be presented for the purpose of calculation of the fractal dimension. Thus, the local degree of `fractality´ is used to differentiate the regions-of-interest (ROI) from the clutter background. Then, the warship is detected based on indiscemibility relation of Rough sets (RS) theory and is recognized based on the geometric characters of warship. Experiments on the image tracker based on TMS320C6416 were conducted, and the results demonstrated that the approach performs well in automatic recognition accuracy and computational efficiency.