
Leather Defect Detection
Software as a Service
NoCategory
Thesis DesignClients
Southern Luzon State University
Techstack
PythonYOLOv5Raspberry PiOpenCVPyQt
Purpose
Southern Luzon State University needed an affordable quality-control aid for leather production — a prototype that automatically spots surface defects so flawed material can be caught before it reaches finished products.
Description
A Raspberry Pi vision prototype that inspects leather surfaces using a custom-trained YOLOv5 model. Camera frames are processed on-device and passed through the detector, which localizes and classifies surface defects in real time, with detections displayed in a PyQt interface. The build includes both detection and non-detection viewing modes and an installation package tailored to the Raspberry Pi, keeping the whole rig compact and low-cost.




