Locating corner points from an edge detected image is very important in view of simplifying the post processing part of a system that utilizes a corner information. In this paper, we propose a robust geometrical approach for corner detection. Unlike classical corner detection methods, which idealize corners as junction points of two line segments, our approach considers the possibility of multiple line segments intersecting at a point. Moreover, junctions caused by two or more curved segments of different curvature are thought of as a corner point. The algorithm has been tested and proved competence with different types of images demonstrating its ability to detect and localize the corners in the image, though we found it to be best suited for images with relatively few curved segments. With the help of non-maximum response suppression technique our approach yields comparatively better result than any other method.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Daniel A. TEFERA, Koichi HARADA, "Geometrical Approach for Corner Detection" in IEICE TRANSACTIONS on Information,
vol. E85-D, no. 4, pp. 727-734, April 2002, doi: .
Abstract: Locating corner points from an edge detected image is very important in view of simplifying the post processing part of a system that utilizes a corner information. In this paper, we propose a robust geometrical approach for corner detection. Unlike classical corner detection methods, which idealize corners as junction points of two line segments, our approach considers the possibility of multiple line segments intersecting at a point. Moreover, junctions caused by two or more curved segments of different curvature are thought of as a corner point. The algorithm has been tested and proved competence with different types of images demonstrating its ability to detect and localize the corners in the image, though we found it to be best suited for images with relatively few curved segments. With the help of non-maximum response suppression technique our approach yields comparatively better result than any other method.
URL: https://global.ieice.org/en_transactions/information/10.1587/e85-d_4_727/_p
Copy
@ARTICLE{e85-d_4_727,
author={Daniel A. TEFERA, Koichi HARADA, },
journal={IEICE TRANSACTIONS on Information},
title={Geometrical Approach for Corner Detection},
year={2002},
volume={E85-D},
number={4},
pages={727-734},
abstract={Locating corner points from an edge detected image is very important in view of simplifying the post processing part of a system that utilizes a corner information. In this paper, we propose a robust geometrical approach for corner detection. Unlike classical corner detection methods, which idealize corners as junction points of two line segments, our approach considers the possibility of multiple line segments intersecting at a point. Moreover, junctions caused by two or more curved segments of different curvature are thought of as a corner point. The algorithm has been tested and proved competence with different types of images demonstrating its ability to detect and localize the corners in the image, though we found it to be best suited for images with relatively few curved segments. With the help of non-maximum response suppression technique our approach yields comparatively better result than any other method.},
keywords={},
doi={},
ISSN={},
month={April},}
Copy
TY - JOUR
TI - Geometrical Approach for Corner Detection
T2 - IEICE TRANSACTIONS on Information
SP - 727
EP - 734
AU - Daniel A. TEFERA
AU - Koichi HARADA
PY - 2002
DO -
JO - IEICE TRANSACTIONS on Information
SN -
VL - E85-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2002
AB - Locating corner points from an edge detected image is very important in view of simplifying the post processing part of a system that utilizes a corner information. In this paper, we propose a robust geometrical approach for corner detection. Unlike classical corner detection methods, which idealize corners as junction points of two line segments, our approach considers the possibility of multiple line segments intersecting at a point. Moreover, junctions caused by two or more curved segments of different curvature are thought of as a corner point. The algorithm has been tested and proved competence with different types of images demonstrating its ability to detect and localize the corners in the image, though we found it to be best suited for images with relatively few curved segments. With the help of non-maximum response suppression technique our approach yields comparatively better result than any other method.
ER -