An ordered list approach to real-time line detection based upon the hough transform

Abstract

This paper presents preliminary results for realtime image processing of video frame rate data. The goal is to identify multiple linear features from video data derived from a camera mounted on a small UAV. The algorithm proposed in this paper adapts the well-known Hough Transform to use sequential image sampling and an ordered list to reduce the computational time and memory allocation required. The results indicate that for images with few lines this algorithm should be able to achieve speeds of better than 25Hz on current desktop computing equipment.