DocumentCode :
262170
Title :
Topological Image Analysis and (Normalised) Representations for Plant Phenotyping
Author :
Janusch, Ines ; Kropatsch, Walter G. ; Busch, Wolfgang
Author_Institution :
Pattern Recognition & Image Process. Group, Vienna Univ. of Technol., Vienna, Austria
fYear :
2014
fDate :
22-25 Sept. 2014
Firstpage :
579
Lastpage :
586
Abstract :
This paper discusses the use of topological image analysis to derive characteristics needed in plant phenotyping. Due to certain features of root systems (deformation over time, overlaps of branches in a 2D image of the root system) a topological analysis is needed to correctly derive these characteristics. The advantages of such a topological analysis are highlighted in this paper and root phenotyping is presented as a new application for computational topology. Characteristics used in plant phenotyping that can be derived from root images using methods of topological image analysis are further presented. A Reeb graph based representation of root images is shown as an example for such a topological analysis. Based on a graph representation a new, normalised representation of root images is introduced.
Keywords :
botany; graph theory; image representation; 2D image; Reeb graph based representation; branch overlapping; computational topology; normalised representation; plant phenotyping; root images; root system features; topological image analysis; Image analysis; Image segmentation; Length measurement; Level measurement; Skeleton; Soil; Topology; Application for Computational Topology; Normalised Root Representation; Reeb Graphs; Root Phenotyping; Topological Image Analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), 2014 16th International Symposium on
Conference_Location :
Timisoara
Print_ISBN :
978-1-4799-8447-3
Type :
conf
DOI :
10.1109/SYNASC.2014.83
Filename :
7034733
Link To Document :
بازگشت