We present a survey on maritime object detection and tracking approaches,which are essential for the development of a navigational system for autonomousships. The electro-optical (EO) sensor considered here is a video camera thatoperates in the visible or the infrared spectra, which conventionallycomplement radar and sonar and have demonstrated effectiveness for situationalawareness at sea has demonstrated its effectiveness over the last few years.This paper provides a comprehensive overview of various approaches of videoprocessing for object detection and tracking in the maritime environment. Wefollow an approach-based taxonomy wherein the advantages and limitations ofeach approach are compared. The object detection system consists of thefollowing modules: horizon detection, static background subtraction andforeground segmentation. Each of these has been studied extensively in maritimesituations and has been shown to be challenging due to the presence ofbackground motion especially due to waves and wakes. The main processesinvolved in object tracking include video frame registration, dynamicbackground subtraction, and the object tracking algorithm itself. Thechallenges for robust tracking arise due to camera motion, dynamic backgroundand low contrast of tracked object, possibly due to environmental degradation.The survey also discusses multisensor approaches and commercial maritimesystems that use EO sensors. The survey also highlights methods from computervision research which hold promise to perform well in maritime EO dataprocessing. Performance of several maritime and computer vision techniques isevaluated on newly proposed Singapore Maritime Dataset.
展开▼
