Deep learning for tooth wear monitoring of mining shovels

The main objective of this project is using deep learning algorithm to enhance the current state of the art tooth wear monitoring system used in mining shovels. Unlike the current approach, the proposed deep learning method operates by building a model from input images in order to make data-driven predictions. We use deep learning approach to identify the pixels that belong to the teeth-line in each video frame taken by camera located on the mining device. Consequently, the extracted teeth will be registered to a template in order to compute all the changes happened to the teeth length during time.

Faculty Supervisor:

Parvin Mousavi

Student:

Sahar Ghavidel

Partner:

Motion Metrics International Corp

Discipline:

Computer science

Sector:

Mining and quarrying

University:

Queen's University