Trichome detection using Machine Learning

2019 Research Internship Project

Faculty Name

Dimitri Androutsos

Project Title

Trichome detection using Machine Learning

Project Description

Trichomes are found on a variety of different plants for a variety of different reasons. Most often they work to protect plants from harmful predators and harsh weather conditions. Some carnivorous plants actually use their trichomes to catch prey because of their stickiness. Animals and other pests are usually deterred from eating or going near cannabis plants because of the bitter taste and potent aroma produced by the trichomes. They also protect the plant, and most importantly its flowers from harsh weather conditions. Trichomes can also inhibit the growth of some fungal diseases. Many types of trichomes are glandular meaning that they produce some kind of secretion or essential oil. Studying the growth rate, size, shape, density and colour of trichomes gives an indication to the maturity of the plant itself and can provide further information such as peak cultivation time and potency. This project will require a student to become familiar with Machine Learning techniques that deal with object recognition (e.g., handwritten characters) and then use that knowledge to do trichome detection from an image database.

Student Responsibility

1) collection of trichome images test database for training purposes via online search sites 2) literature review 3) learn about Machine Learning (ML) techniques for object identification 4) implement an ML system that performs trichome detection

Specific Requirements

Student must know and be comfortable to program (a real programming language, not just MATLAB) and be ready to learn new programming tools to perform Machine Learning. Student, ideally would know some rudimentary image processing, but that is not essential.

Reseach Internship Application

Dimitri Androutsos : Trichome detection using Machine Learning | Wednesday April 3rd 2019 03:00 PM