Predicting plant environmental exposure using remote sensing
File(s)
Author(s)
Adams, Christopher
Type
Thesis
Abstract
Wheat is one of the most important crops globally with 776.4 million tonnes produced in
2019 alone. However, 10% of all wheat yield is predicted to be lost to Septoria Tritici
Blotch (STB) caused by Zymoseptoria tritici (Z. tritici). Throughout Europe farmers spend
£0.9 billion annually on preventative fungicide regimes to protect wheat against Z. tritici. A
preventative fungicide regime is used as Z. tritici has a 9-16 day asymptomatic latent phase
which makes it difficult to detect before symptoms develop, after which point fungicide
intervention is ineffective.
In the second chapter of my thesis I use hyperspectral sensing and imaging techniques,
analysed with machine learning to detect and predict symptomatic Z. tritici infection in
winter wheat, in UK based field trials, with high accuracy. This has the potential to
improve detection and monitoring of symptomatic Z. tritici infection and could facilitate
precision agriculture methods, to use in the subsequent growing season, that optimise
fungicide use and increase yield.
In the third chapter of my thesis, I develop a multispectral imaging system which can detect
and utilise none visible shifts in plant leaf reflectance to distinguish plants based on the
nitrogen source applied. Currently, plants are treated with nitrogen sources to increase
growth and yield, the most common being calcium ammonium nitrate. However, some
nitrogen sources are used in illicit activities. Ammonium nitrate is used in explosive
manufacture and ammonium sulphate in the cultivation and extraction of the narcotic
cocaine from Erythroxylum spp. In my third chapter I show that hyperspectral sensing,
multispectral imaging, and machine learning image analysis can be used to visualise and
differentiate plants exposed to different nefarious nitrogen sources. Metabolomic analysis
of leaves from plants exposed to different nitrogen sources reveals shifts in colourful
metabolites that may contribute to altered reflectance signatures. This suggests that
different nitrogen feeding regimes alter plant secondary metabolism leading to changes in
plant leaf reflectance detectable via machine learning of multispectral data but not the
naked eye. These results could facilitate the development of technologies to monitor illegal
activities involving various nitrogen sources and further inform nitrogen application
requirements in agriculture.
In my fourth chapter I implement and adapt the hyperspectral sensing, multispectral
imaging and machine learning image analysis developed in the third chapter to detect
asymptomatic (and symptomatic) Z. tritici infection in winter wheat, in UK based field
trials, with high accuracy. This has the potential to improve detection and monitoring of all
stages of Z. tritici infection and could facilitate precision agriculture methods to be used
during the current growing season that optimise fungicide use and increase yield.
2019 alone. However, 10% of all wheat yield is predicted to be lost to Septoria Tritici
Blotch (STB) caused by Zymoseptoria tritici (Z. tritici). Throughout Europe farmers spend
£0.9 billion annually on preventative fungicide regimes to protect wheat against Z. tritici. A
preventative fungicide regime is used as Z. tritici has a 9-16 day asymptomatic latent phase
which makes it difficult to detect before symptoms develop, after which point fungicide
intervention is ineffective.
In the second chapter of my thesis I use hyperspectral sensing and imaging techniques,
analysed with machine learning to detect and predict symptomatic Z. tritici infection in
winter wheat, in UK based field trials, with high accuracy. This has the potential to
improve detection and monitoring of symptomatic Z. tritici infection and could facilitate
precision agriculture methods, to use in the subsequent growing season, that optimise
fungicide use and increase yield.
In the third chapter of my thesis, I develop a multispectral imaging system which can detect
and utilise none visible shifts in plant leaf reflectance to distinguish plants based on the
nitrogen source applied. Currently, plants are treated with nitrogen sources to increase
growth and yield, the most common being calcium ammonium nitrate. However, some
nitrogen sources are used in illicit activities. Ammonium nitrate is used in explosive
manufacture and ammonium sulphate in the cultivation and extraction of the narcotic
cocaine from Erythroxylum spp. In my third chapter I show that hyperspectral sensing,
multispectral imaging, and machine learning image analysis can be used to visualise and
differentiate plants exposed to different nefarious nitrogen sources. Metabolomic analysis
of leaves from plants exposed to different nitrogen sources reveals shifts in colourful
metabolites that may contribute to altered reflectance signatures. This suggests that
different nitrogen feeding regimes alter plant secondary metabolism leading to changes in
plant leaf reflectance detectable via machine learning of multispectral data but not the
naked eye. These results could facilitate the development of technologies to monitor illegal
activities involving various nitrogen sources and further inform nitrogen application
requirements in agriculture.
In my fourth chapter I implement and adapt the hyperspectral sensing, multispectral
imaging and machine learning image analysis developed in the third chapter to detect
asymptomatic (and symptomatic) Z. tritici infection in winter wheat, in UK based field
trials, with high accuracy. This has the potential to improve detection and monitoring of all
stages of Z. tritici infection and could facilitate precision agriculture methods to be used
during the current growing season that optimise fungicide use and increase yield.
Version
Open Access
Date Issued
2020-09
Date Awarded
2021-03
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Windram, Oliver
Sponsor
Biotechnology and Biological Sciences Research Council (Great Britain)
The Douglas Bomford Trust (DBT)
Publisher Department
Department of Life Sciences (Silwood Park)
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)