Determination of length and width of a line-segment by using a hough transform


by Z Xu, B-S Shin, R Klette
Abstract:
The standard Hough transform does not provide length and width of a line-segment detected in an image; it just detects the normal parameters of the line. We present a novel method for determining also length and width of a line segment by using the Hough transform. Our method uses statistical analysis of voting cells around a peak in the Hough space. In image space, the voting cells and voting values are analysed. The functional relationship between voting variance and voting angle is deduced. We approximate this relationship by a quadratic polynomial curve. In Hough space, the statistical variances of columns around a peak are computed and used to fit a quadratic polynomial function. The length and width of a line segment are determined simultaneously by resolving the equations generated by comparing the corresponding coefficients of two functions. We tested and verified the proposed method on simulated and real-world images. Obtained experimental results demonstrate the accuracy of our novel method for determining length and width of detected line segments.
Reference:
Determination of length and width of a line-segment by using a hough transform (Z Xu, B-S Shin, R Klette), In Discrete Gemoetry for Computer Imagery, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Springer Verlag, volume 8668, 2014.
Bibtex Entry:
@inproceedings{xu2014determinationtransform,
author = "Xu, Z and Shin, B-S and Klette, R",
booktitle = "Discrete Gemoetry for Computer Imagery, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
organization = "Siena, Italy",
pages = "190--201",
publisher = "Springer Verlag",
title = "Determination of length and width of a line-segment by using a hough transform",
volume = "8668",
year = "2014",
abstract = "The standard Hough transform does not provide length and width of a line-segment detected in an image; it just detects the normal parameters of the line. We present a novel method for determining also length and width of a line segment by using the Hough transform. Our method uses statistical analysis of voting cells around a peak in the Hough space. In image space, the voting cells and voting values are analysed. The functional relationship between voting variance and voting angle is deduced. We approximate this relationship by a quadratic polynomial curve. In Hough space, the statistical variances of columns around a peak are computed and used to fit a quadratic polynomial function. The length and width of a line segment are determined simultaneously by resolving the equations generated by comparing the corresponding coefficients of two functions. We tested and verified the proposed method on simulated and real-world images. Obtained experimental results demonstrate the accuracy of our novel method for determining length and width of detected line segments.",
startyear = "2014",
startmonth = "Sep",
startday = "10",
finishyear = "2014",
finishmonth = "Sep",
finishday = "12",
issn = "0302-9743",
eissn = "1611-3349",
keyword = "Curve fitting",
keyword = "Hough transform",
keyword = "Length",
keyword = "Width",
language = "eng",
conference = "18th IAPR International Conference, DGCI",
}