A simple tracking approach for line extraction
Liang Chen, Lei Guo, Yaping Wang, Shiwei Gao, Yong Yu · 2007
Line extraction plays an important role in many applications as mid-level feature description; the problem is how to get accurate location of line effectively. In this paper, the extraction is simplified by edge tracking with adding some constraints which ensure the edge be linked near to the line. It is a strategy that made a feasible way to obtain lines. Following edge detecting, the proposed approach obtains lines by tracking edge with priority direction, inertia tracking, judging and rectifying linear relation real-time. After such constraints on the tracking process, the edge is linked to a line. Experiments show varied results of proposed approach and the contrast to Hough transform, and it can be seen that the line extracted is more effective than latter.