Warehouse Computer Vision System

Real-time forklift and pallet tracking system using YOLOv8 + ByteTrack with QR code identification (pyzbar). Supports RTSP, local video, and webcam feeds. Generates CSV movement logs for WMS integration.

Overview

A production-ready warehouse computer vision system for real-time tracking of forklifts and pallets across warehouse environments. Built to integrate directly with Warehouse Management Systems (WMS) via CSV movement logs.

Features

  • Multi-object tracking — YOLOv8 detection + ByteTrack for persistent object IDs across frames
  • QR code identification — pyzbar integration for reading pallet and forklift QR codes in motion
  • Flexible input — supports RTSP camera streams, local video files, and webcam
  • WMS-ready output — CSV logs of object ID, location, timestamp, and movement events for direct WMS ingestion
  • Real-time visualization — annotated video feed with bounding boxes, track IDs, and QR data overlay

Architecture

Camera Feed (RTSP/file/webcam)
        ↓
YOLOv8 Object Detection (forklift, pallet classes)
        ↓
ByteTrack Multi-Object Tracker (persistent IDs)
        ↓
pyzbar QR Decoder (identity resolution)
        ↓
Movement Event Logger (CSV) → WMS Integration

Use Cases

  • Inventory movement tracking without manual scanning
  • Forklift path analysis for safety compliance
  • Real-time zone occupancy for dynamic slotting
  • Audit trail generation for warehouse operations

Tech Stack

Python · YOLOv8 (Ultralytics) · ByteTrack · pyzbar · OpenCV · NumPy · Pandas