Line detection in digital pictures: A hypothesis prediction/verification paradigm

Automatic extraction of parallel edges based on eigenvalue analysis and collateral expansion

“…Canny 1986, Callaghan et al 2003, Karantzalos and Argialas 2006, McLaughlin 2006, Wei and Bartels 2006, these methods mostly originated from, and until now have still focused on, single edges. For theoretical expansion, several models have been added together, such as the statistics-based category (Mansouri et al 1987), the gradient-based category (Nelson 1994) and the Hough transform (HT)-based category ( Hart 1972). Nevertheless, these models commonly take the mathematical singularity of boundaries as their theoretical basis, and are also required to be all-prevalent.…”

A Real-Time Hough-Based Method for Segment Detection in Complex Multisensor Images

“…The former is more time-consuming, but allows one to achieve higher line orientation and localization accuracies. Image-space methods include edge-tracking and contour-following algorithms [1], chain codes [2], prediction/veri®cation mechanisms [3] and relaxation algorithms [4]. One of the most signi®cant contributions has been carried out by Burns et al [5] who developed a method for extracting segments from images by using available information about the region around a straight line.…”

A simple and robust line detection algorithm based on small eigenvalue analysis

“…Over decades, several models have been reported in literature and are broadly classified into four categories viz., statistical based (Mansouri et al, 1987), gradient based (Burns et al, 1986;Nelson, 1994), pixel connectivity-edge linking based (Nevatia and Babu, 1980) and Hough transform (HT) based (Duda and Hart, 1972) models. Mansouri et al (1987) proposed a hypothesizeand-test algorithm to extract line segments of specified lengths by hypothesizing their existence by the use of local information. In their attempt, the statistical properties of a digital model of an ideal segment were explored to verify the hypothesis and thus, it is generally classified as a statistical based approach.…”