Sea-Sky Line Extraction with Linear Fitting Based on Line Segment Detection

Sea-sky line extraction under complex sea-sky background plays a very important role in image processing for the detection and tracking of targets. In this paper, different from traditional methods, we proposed a novel sea-sky line extraction approach with straight line fitting based on line segment detection. Firstly, a fast and robust line segment detection method was used to locate the sea-sky region roughly. In this algorithm, the line segment region detection can be completed in linear time. Then, the accurate parameters of sea-sky line is computed in an improved linear fitting method based on least squares. The experimental results indicate that the proposed approach can effectively extract the sea-sky line from sea-sky background image.