Generalised linear model and analysis of cereal plant biomass

Date

2013

Authors

Cespedes, M.
Cai, J.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Conference paper

Citation

Proceedings - 20th International Congress on Modelling and Simulation, MODSIM 2013, 2013, pp.503-509

Statement of Responsibility

Conference Name

20th International Congress on Modelling and Simulation, MODSIM 2013 (1 Dec 2013 - 6 Dec 2013 : Adelaide, Australia)

Abstract

Numerous literature discuss methods, models and results achieved by applied image analysis techniques in order to deduce plant information. Although all methods required data from destructive testing their recommendations suggest the expansion of their methods onto an automated system is within reach. This new system will provide a quicker, cost efficient non-intrusive way to gather plant information as an alternative to destructive testing. The estimation of plant biomass is important to many applications such as plant breeding and agriculture. The generic biomass model presented here is based on the separation of plant components which opens up the potential to account for the structural and density differences of plant components. Various statistical techniques have been employed to assess the improvement of previous biomass models and challenges which arise when regression models do not conform to assumptions. This paper offers an alternative model which preserves the original data and linear model template, yet accounts for increasing systematic variability present in the data. Both informal and formal statistical methods were used to asses the goodness of fit and comparison between the commonly used method of least squares regression and the generalised model proposed here. These models are an extension to those found in literature of plants under experimental conditions and the intention is to apply this model coupled with image analysis in order to deduce plant information in a non-intrusive way. The proposed biomass models presented here aim to combine image processing techniques with mathematical modeling in an effort to replace destructive testing

School/Discipline

Dissertation Note

Provenance

Description

Access Status

Rights

Copyright 2020 The author(s)

License

Grant ID

Published Version

Call number

Persistent link to this record