EasyAI Produce Grading Prototype
An early prototype of an image-AI produce quality prediction system
EasyAI·2020.12 - 2021.04·Software development
Python
OpenCV
PyTorch
Background
At EasyAI, this project produced an early prototype of an image-AI system for predicting produce quality. Grading by eye on the sorting floor was inconsistent and limited in throughput, so the goal was automated quality classification based on image analysis.
System architecture
How it was built
I built an image preprocessing pipeline on Python and OpenCV, then implemented the produce quality prediction logic with an AI model. Designing the core data flow of the sorting system at this stage established the structure that later industrial systems were built on.
Results
- Completed a prototype of an AI-based produce quality prediction system.
- When the EasyAI AI sorting division spun off as AIOFARM, this prototype became the starting point for the sorting system backend, the GPU vision pipeline, and the desktop client.