Proposal of new Hough transform algorithm using robust statistics: LMedS Hough transform

摘要

In this paper, we propose a new Rough transform algorithm called LMedS Rough transform, which uses the measure of least median of squares as the basis to estimate the line. This means that LMedS Rough transform can provide a new measure beyond the majority standard of the ordinary Rough transform, and therefore that LMedS Rough can detect lines just in the same way of LMedS line fitting procedure. In addition to this, because this algorithm is constructed on the Rough transform paradigm, the basic useful properties of Rough transform such as robustness to the noise, multi-line detection and global line detection are inherited in LMedS Rough algorithm.