Efficient Horizon Detection On Complex Sea For Sea Surveillance
Obstacle detection is one of the key features of an unmanned surface vehicle (USV). Research on real time obstacle
detection gains momentum these days. Horizon detection is one of the fundamental support factors when performing real-time
obstacle detection. Precise detection helps to improve the efficiency and accuracy of obstacle detection becausethis eliminates
irrelevant areas as much as possible. In this paper, three different algorithms are applied on three different sequences of images
that were taken on the open sea. This paper focuses on the success rate and time efficiency of horizon detection. By applying
four different algorithms, includingProbabilistic Hough Transform, Least Square Optimization, Random Sample Consensus
(RANSAC) and Generic line detection method, we extract the horizon from the images. The experimental results show that
different methodshave different time efficiency and accuracy for horizon detection. By comparing the advantages and
disadvantages of each algorithm, it is possible to determine which algorithm is more suitable under given constraints.
Index Terms- Horizon detection, Image rectification, Sea surface estimation, Sea surveillance.