A Fast Sea-sky Line Detection Method Based on EDLines
Yongqing Zhang, Huigang Wang, Runhe Yao · 2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) · 2021
As the dividing line between the sky area and the water surface area, sea-sky line (SSL) can not only be used to extract the area where the target may exist, but also can be used to stabilize the image collected on the sailing unmanned surface vehicle (USV). The image obtained by the ship-borne camera contains a lot of interference due to factors such as waves and reflections. Existing methods cannot guarantee the accuracy, robustness, and real-time requirements of SSL detection at the same time. In our method, the pixels where the gradient is a local peak in the vertical direction are chosen as the key points of detection and a novel and effective pixel connection rule is proposed. The proposed method is tested on visual images collected by USV. While ensuring accuracy and robustness, the average processing time of a single frame of image is less than that of the existing methods and meets real-time requirement.