Enhancement and in-field deployment of spectroscopy-based nutrient sensor for potato plants using machine learning and IoT techniques

This is a project that aims to deploy a new sensing system that allows the estimate of nutrients in potato plants in near real-time by scanning their leaves in two different modes: fresh when still intact and dried. This sensing system would replace manual sampling and wet tissue chemical analysis to allow immediate response to nutrient deficiency that will optimize the use of fertilizers and contribute to reducing greenhouse gas emission – in addition to the economic benefits of using fertilizers only as needed. In order to be deployed, the sensing system should be validated and tested which this project will help to achieve. This step is critical because the sensing system relies on machine learning whose performance improves as more data are used for training and evaluation. Hence, maximizing the amount of data is one of the objectives of this project. Also, the enhanced machine learning models will be plugged in a computational cloud around which wireless connectivity will be developed which is the other objective in this project.

Faculty Supervisor:

Ahmad Al-Mallahi

Student:

Partner:

McCain Foods

Discipline:

Engineering

Sector:

Manufacturing

University:

Dalhousie University

Program: