Search Aid System Based on Machine Vision and Its Visual Attention Model for Rescue Target Detection

The prompt search and rescue of lifesaving target is very important in the case that a marine casualty occurs. To detect the small target in the wide views over the sea, we have proposed a machine vision system to aid search and rescue on the sea, which combines remote sensing, radar, infrared with visual light technology. One of the detection methods in this system, which is based on visual attention mechanism, is proposed in this paper to find rescue target from visual light image. The color information, image intensity and other image properties are used to generate the feature maps that form the saliency map subsequently with the weighted integrating strategy. The experimental results show that the proposed method is efficient to detect the small target in cluttered ocean scene.