@inproceedings{8715986,
  abstract     = {{Crop diseases remain one of the key yield-limiting factors in leek production, causing the need for continuous fungicide applications, resulting in socio-economic and environmental costs. Precision agriculture aims at reducing these costs, while maintaining or improving farmer revenues. The first requirement for the implementation of variable rate fungicide application is in-field measurement of the spread of disease. In this work, a disease detection methodology was created that could detect early stages of disease development in field conditions for two main diseases: leek rust and white tip disease. A hyperspectral training library was constructed for both diseases separately, containing spectra corresponding to healthy leek, weed plants, diseased leek plants and soil. A custom preprocessing and soil pixel removal strategy was constructed for each disease. An evaluation of 11 common classifiers was performed, of which a logistic regression classifier provided the highest classification accuracy. This logistic regression supervised machine learning classifier was then trained on the hyperspectral library. For leek rust disease, the focus was on detecting early infection (i.e. single rust pustules). For white tip disease, the model was used to classify the data into four classes: healthy plant material, early (pre-visual) disease, moderate disease, severe disease and fully developed disease. The overall accuracy of the disease model was 98.14% for rust and 96.74% for white tip disease. It can be concluded that the results in this work are an important step towards the mapping of leek rust and white tip disease, and that future research is needed to overcome certain challenges before variable rate fungicide applications can be adopted against leek diseases.template is designed to assist a high quality proceedings book editing and printing.}},
  articleno    = {{ID 4603}},
  author       = {{Appeltans, Simon and Pieters, Jan and Mouazen, Abdul}},
  booktitle    = {{EurAgEng 2021 Conference, Proceedings}},
  keywords     = {{Hyperspectral,Disease Detection,Leek White Tip Disease,Leek Rust Disease,Machine learning}},
  language     = {{eng}},
  location     = {{Online (Evora, Portugal)}},
  pages        = {{9}},
  title        = {{Detection of leek rust and white tip disease under field conditions using hyperspectral proximal sensing and supervised machine learning}},
  url          = {{https://ebook.ajnet.net/wp-content/uploads/2021/06/Abreu-Livro-de-Resumos-AGENG2021-Formatacao-v8-3-de-julho.pdf}},
  year         = {{2021}},
}