Proof of concept for automated mushroom harvesting

Mushrooms for the retail environment are grown in climate-controlled indoor environments. The crop develops very quickly and must be harvested over a period of three or four days. Harvest yield depends to a large extent on the skill of the harvester; if the right mushroom is picked at the right time, a new mushroom can grow in the same place, yielding up to three generations of mushroom from a single seeding. In this project, we will investigate technologies for identifying and gripping mushrooms as a first step towards the development of an automated system for picking mushrooms at the optimal size and ripeness for picking. We will develop a novel gripper design, most likely based on smart materials, and a sensor fusion-based approach to mushroom recognition, using some combination of visual, thermal and hyperspectral imagers.

This project, which provides the seed for a series of largerscale collaborative projects over the next five years, has as its long term goal the development of an automated mushroom harvester that should lead to as much as a 25% increase in harvest efficiency. We will make the solution available to the industrial partner and the wider farming community internationally.

Faculty Supervisor:

Drs. Kenneth A. McIsaac & Mehrdad Kermani

Student:

Payman Yadmellat & Alireza Masoudian

Partner:

Whitecrest Mushrooms

Discipline:

Engineering - computer / electrical

Sector:

Agriculture

University:

Western University