A sensor for vision-based navigation in underwater path tracking with color and edge segmentation

Abstract

This paper aims the design and implementation of a visionbased sensor for navigation of underwater vehicles with adaptive attributes. The objective pointed out is a sensor for tracking of underwater lines. The sensor employs a basic structure with a pixel-wise AND operation of binarized frames of separated channels HSV and an edgesegmented frame. The basic sensor performs well by good illuminated scenes. By significant drops of luminance, the efficiency falls. So an adaptive sensor is proposed over the basic structure. It operates on the brightness channel carrying out a maximization of contains in the accumulator bins of a Hough transformation. It has proven to enhanced the identification of the tracked line increasing the success rate.