Research on a Horizon Line Detection Method for Unmanned Surface Vehicles in Complex Environments

A critical step in the visual navigation of unmanned surface vehicles (USVs) is horizon line detection, which can be used to adjust altitude as well for obstacle avoidance complex environments. In this paper, a real-time and accurate detection method proposed. Our approach first differentiates complexity navigational scenes using angular second moment (ASM) parameters grey level co-occurrence matrix (GLCM). Then, region interest (ROI) initially extracted minimal human interaction scenes, while subsequent frames are dynamically acquired automatic feature point matching. The matched ROI maximally removed from background, Zernike-moment-based edges obtained ROI. Finally, complete sea information through linear fitting lower edge points information. Through various experiments carried out on classical dataset, our own datasets, that another previously published we illustrate significance accuracy technique results show performance has potential applications autonomous control USVs.