对地观测系统可以快速检测大面积区域, 为世界各沿海国家的船只监测任务带来了极大便利。针对遥感影像中的船只自动化检测问题, 利用船只相对于海面呈现明显的视觉显著特性, 提出了一种基于船只与海面反差特性的船只检测方法, 将对船只的检测定位转变为对场景中显著目标的发现与分割问题。方法以视网膜到初级视觉皮层V1这一视觉通路的生理过程为指导, 模拟了其中视觉信号的产生机制以获得场景的显著图。以光学影像为例进行的实验结果显示, 在没有目标先验知识的前提下, 基于反差特性的方法能有效地将视觉注意力集中于海面船只目标, 取得了较好的检测结果。

Earth observation systems can quickly detect objects in a large area, which offers great convenience for ship detection and monitoring by the coastal countries in the world. Towards the automated ship detection in remote sensing images, a method that made use of the salient contrast between ship and sea was proposed. Thus, the ship detection and segmentation problem was converted to a salient object discovery and location problem. Inspired by the biological process of visual pathway from the retina to the primary visual cortex V1, the visual signal generating mechanism was simulated to acquire the saliency map of the scene. The experimental results, obtained from using optical images, show that the contrast based method can effectively concentrate on the ships in the sea without any prior knowledge. It is noteworthy that the proposed method achieves good results.