21734 articles – 15570 references  [version française]
HAL: hal-00651251, version 1

Detailed view  Export this paper
Available versions:
Texture Analysis with Arbitrarily Oriented Morphological Opening and Closing
Jan Bartovsky 1, Eva Dokladalova 2, Michel Bilodeau 1, Petr Dokládal ( ) 1
(2011-12-13)

This paper presents a fast, streaming algorithm for 1-D morphological opening on 2-D support. The algorithm is further extended to compute the complete size distribution during a single image run. The Structuring Element (SE) can be oriented under arbitrary angle that allows us to perform different orientation-involved image analysis, such as local angle extraction, directional granulometries, \etc The algorithm processes an image in constant time irrespective of the SE orientation and size, with a minimal latency and very low memory requirements. Regardless the SE orientation, it reads and writes data strictly sequentially in the horizontal scan order. Aforementioned properties allow an efficient implementation in embedded hardware platforms that opens a new opportunity of a parallel computation, and consequently, a significant speed-up.
1:  Centre de Morphologie Mathématique (CMM)
MINES ParisTech - École nationale supérieure des mines de Paris
2:  Laboratoire d'Informatique Gaspard-Monge (LIGM)
Université Paris-Est Marne-la-Vallée (UPEMLV) – ESIEE – Ecole des Ponts ParisTech – Fédération de Recherche Bézout – CNRS : UMR8049
Mathematics

Computer Science
Mathematical Morphology – Granulometry – parallel implementation
Attached file list to this document: 
PDF
main.pdf(466.4 KB)