All projects
Project 07
Bagasort
A computer-vision sorting machine for oranges, built as the capstone project of my Computer Engineering degree.
- Status
- In development
- Role
- Capstone project for the B.Sc. in Computer Engineering.
- Source
- Private codebase
- Cameras
- Raspberry Pi (CNN)
- ESP32 firmware
- Sorting mechanism
Problem
Sorting fruit by quality by hand is slow and inconsistent. The project explores whether a low-cost machine with cameras and an embedded model can classify and sort oranges reliably on its own.
Solution
A Raspberry Pi running the Python core with two cameras, a convolutional neural network trained with TensorFlow and deployed for on-device inference, ESP32 firmware driving the mechanical parts, and a web interface for operation and monitoring.
Engineering highlights
- Two-camera capture for a fuller view of each fruit.
- CNN trained with transfer learning and deployed for edge inference.
- Firmware built with PlatformIO, coordinating with the Python core.
- React and Tailwind interface for operating the machine.