A Method of Ship Detection in Optical Satellite Image Based on Saliency Map
Zehui Li, Xiaofang Xie, Weihua Zhao, Yong Liu · 2013
Ship detection from space is an important application for maritime traffic surveillance. However, fewer efforts were devoted to research on automatic detection from optical satellite imagery. This paper presents current progress in ship detection from optical imagery and provides an automatic ship detection algorithm based on the combination of a random saliency map and an image pyramid to detect vessels in ports or straits. The algorithm was tested with an average detectability of 83.2% for large and medium vessels and false alarm rate of 33.5%.