Straight Line Extraction Using Iterative Total Least Squares Methods

Visual Communication and Image Representation, 6(1) 59-68, 1993

10 Pages Posted: 27 Apr 2014

Date Written: August 11, 1993

Abstract

In this paper we present a new algorithm for enhancing the accuracy of the parameter extraction of straight lines in a two-dimensional image. The algorithm achieves high accuracy in comparatively less computational time than most traditional methods and is invariant under rotation and translation. The Iterative Total Least Squares (ITLS) method starts from an initial estimate of the line parameters. When no a priori information about the image is available this estimate can be assigned randomly. Alternately, a lower accuracy method can be used to generate an initial estimate which will result in faster convergence. Then, a rectangular window is centered using the current line approximation, and a new line estimate is generated by making a total least squares fit through the pixels contained within the window. This is repeated until convergence is reached. Adaptively adjusting the window size yields the 4D ITLS process. In addition, a pairwise accelerated ITLS method has been developed which substantially increases the convergence rate. We conclude with some examples where the ITLS method has been used successfully.

Suggested Citation

Van Mieghem, Jan Albert and Avi-Itzhak, Hadar and Melen, Roger, Straight Line Extraction Using Iterative Total Least Squares Methods (August 11, 1993). Visual Communication and Image Representation, 6(1) 59-68, 1993, Available at SSRN: https://ssrn.com/abstract=2336194

Jan Albert Van Mieghem (Contact Author)

Northwestern University - Kellogg School of Management ( email )

2001 Sheridan Road
Evanston, IL 60208
United States

Hadar Avi-Itzhak

Canon Research Center America, Inc

4009 Miranda Avenue
Palo Alto, CA 94304
United States

Roger Melen

affiliation not provided to SSRN

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
60
Abstract Views
975
Rank
643,103
PlumX Metrics