Part Number Smart Sorter (Computer Vision)

Overview

Engineered an industrial Computer Vision Cover Part Number Smart Sorter for PT. Century Batteries Indonesia (Astra Otoparts). Utilizing a custom-trained YOLO object detection model achieving ~99% mAP50, the system automates battery cover part number verification on the assembly line.

Key Impacts & Highlights

Efficiency: Cut per-tray inspection time by 70%.
Accuracy: Reduced misidentification error rate from ±5-10% down to <1%.
Cost Reduction: Eliminated ~Rp100,000,000/year in manual inspection and sorting labor costs.
Tech Stack: Python, PyTorch, YOLOv8, OpenCV, Docker.

Demo

Classification Method Demo

HPVT

Project Description

General Zoom

Archi Zoom

General Zoom

Overview

Key Impacts & Highlights

Efficiency: Cut per-tray inspection time by 70%.
Accuracy: Reduced misidentification error rate from ±5-10% down to <1%.
Cost Reduction: Eliminated ~Rp100,000,000/year in manual inspection and sorting labor costs.
Tech Stack: Python, PyTorch, YOLOv8, OpenCV, Docker.

Demo

Classification Method Demo

HPVT

Project Description

General Zoom

Archi Zoom

General Zoom