A Multi-Scale Fractal Dimension based Onboard Ship Saliency Detection Algorithm

Detection of ship targets in the sea area is an important field in remote sensing image target detection. As the ships and the surrounding areas are very different in texture, that makes it a possible solution to detect the ships using the texture feature. Aiming at the detection of ship targets, a novel ship target detection algorithm in a large scene of the optical remote sensing image is proposed in this paper. This algorithm is based on the conspicuity of ship targets of multi-scale fractal dimension feature in the sea background, and then the detection of ship targets is realized by the method of visual saliency model. In this paper, the accuracy of fractal dimension feature of small or medium-sized window by using differential box counting algorithm has been improved. The novel algorithm proposed in this paper is based on the significant difference of natural background and man-made objects in multi-scale fractal dimension feature. Then, the conspicuous fractal feature is obtained by using center-surround difference arithmetic operator, in order to highlight the target in the saliency map normalization is need in the final step. On the basis of the saliency map the rapid detection of ship targets in the sea background can be realized. Experimental results show that ship targets in the sea background can be detected accurately with this algorithm, and also the false alarm rate has been effectively reduced.