A Saliency Gradient Feature Based Sea-Sky-Line Detection Algorithm for USVs Navigation

摘要

The sea-sky-line detection plays an important role for USVs environmental conception and their navigation. To reduce the negative affection from the complex background and USVs motion, the saliency gradient feature is firstly proposed for feature extraction. Then Hough line detection is applied. By result marking, we obtain the position of the sea-sky-line in the original image finally. The experimental results prove that the proposed algorithm performs stably at high detection rate in different complex environment containing wave, rotation and weather condition factors. This method is suitable for the sea-sky-line detection of USVs navigation tasks.