Automated pest detection in plant trials
Project Overview
With efforts to improve the efficiency of companion planting where plants are grown alongside trial plants to attract predators that eat pests, we develop a machine learning model to detect these pests and track their numbers across the growing season.
What were the key results of your research project?
- Worked on a pipeline that properly cleans and refines the data, establishing the foundation to develop a reliable model.
- Solved the class imbalance problem by randomly splitting the images into train/test before augmenting insect samples.
- Developed a classification model with about 85% accuracy.
- Implemented a clustering algorithm to solve the problem of multiple volunteers clicking on the same bug, to lay the groundwork for the next steps of this project.
Github repository: https://github.com/RHS-Wisley-Bug-Watch/RHS-Pests
A presentation of this research will be shared here and on our YouTube site when available.
How do you feel you have benefitted from completing this internship and has it made you consider future career paths?
The internship has given me the opportunity to explore my interests in research software engineering and machine learning.
I've managed to develop my technical and research skills, and expose myself to new experiences such as using the High-End Compute service and manipulating images. I have also honed my ability to explain technical details in a simpler way.
I thoroughly enjoyed my experience and I look forward to doing more in research software engineering.
Download presentation slides