Reducing crop loss and increasing income is one main objective of farmers. Rapid identification of plant diseases, which stifle plant development, is a challenge faced by farmers. Deep learning has shown great potential in the field of image recognition and anomaly detection.
A serious threat to food security is crop diseases. Detecting crop diseases in the early stage is vital for farmers to treat and revive such plants thus increasing crop produce and income. However, it is challenging for farmers to detect crops with diseases in the early stage. We review prior studies and analyse existing approaches used in detecting crop diseases. Next, a large comprehensive study is conducted and various deep learning approaches are compared and evaluated. Finally, a novel machine learning/deep learning approach is designed to aid improve disease detection.
- Review previous work and datasets on the application of machine/deep learning approaches in plant disease detection
- Conduct large scale comprehensive empirical experiments with already existing datasets using Python/R (e.g. pyTorch, Tensorflow, sci-kit learn) (support with the coding can be provided by the supervisor)
- Design a novel approach to improve disease detection performance
The work in this master thesis entails:
- To collect full-text articles or PDFs from primary studies and SLRs in the plant/crop disease field
- To assess the challenges and solutions available in the literature/ in practice
- To design and develop a Deep Learning algorithm that improves the detection performance of current state-of-the-art models
- Mohanty, Sharada P., David P. Hughes, and Marcel Salathé. "Using deep learning for image-based plant disease detection." Frontiers in plant science 7 (2016): 1419.- https://www.frontiersin.org/articles/10.3389/fpls.2016.01419/full
- Courses: Programming in Python (INF-22306), Artificial Intelligence (INF-5006) (Optional), Data Science Concepts (INF-34306) or Machine Learning (FTE-35306)
- Required skills/knowledge: basic data analytics and willingness to learn new software tools, interest about plants and crops
Key words: Deep Learning, Software engineering