Rectilinear structure extraction in textured images with an irregular, graph-based Markov random field model

Author

Delagnes, Philippe ; Barba, Dominique

Author_Institution

IRESTE, SEI Lab. EP CNRS, Nantes, France

Volume

2

fYear

1996

fDate

25-29 Aug 1996

Firstpage

800

Abstract

This paper presents the application of Markov random field modelling to the extraction of poorly contrasted linear structures in textured areas. After a line feature detection step is performed, a set of straight line segments is derived from the feature image. This set of segments constitutes an irregular lattice which reproduces the image rectilinear pattern with accuracy. A Markov random field is then defined on this lattice, in order to group the sites that belong to the same structure. Finally the Markovian segmentation can be post-processed in order to extract global patterns. Results are given on pavement distress images

Keywords

Markov processes; computer vision; edge detection; feature extraction; image segmentation; image texture; structural engineering computing; Markov random field model; irregular lattice; line feature detection; pavement distress images; rectilinear structure extraction; segmentation; straight line segments; textured images; Collaborative work; Computer vision; Data mining; Electronic mail; Image segmentation; Image texture analysis; Laboratories; Lattices; Layout; Markov random fields;

fLanguage

English

Publisher

ieee

Conference_Titel

Pattern Recognition, 1996., Proceedings of the 13th International Conference on

Conference_Location

Vienna

ISSN

1051-4651

Print_ISBN

0-8186-7282-X

Type

conf

DOI

10.1109/ICPR.1996.547186

Filename

547186