Apple Grade Classifier
Machine learning system for automated apple grading using computer vision and multi-camera analysis.
This innovative project focuses on the development of an apple classifier machine based on apple grades and their defects. The system revolutionizes traditional manual sorting methods through automated quality assessment.
The project entails a machine learning classifier (binary and multi-tiered) using multiple camera angles along a "conveyor belt" style process, where apples are fed in and the system automatically detects issues with the presented apple. This is preferred over manual sorting as a result of labour expense and inefficiency.
Aside from apple grades, which are based on the BC Tree Fruits grades table, the machine involves sugar and starch level detection and furthermore an electromechanical system which will automatically move "bad" apples away and let "good" apples continue.
My role specifically focused on the development of the machine learning model and developing robust methods for testing sugar and starch contents.