v0.1 – Prototype

v0.1 – Prototype

First prototype of a simple YOLOv8-based object detection tool.

This version is intentionally minimal: one script (detect.py), one batch helper (start.bat), and a tiny requirements.txt. It’s meant as a clean starting point, not a polished product.

---

✨ What’s in this release

• 🔍 Real-time object detection

  • Runs YOLOv8 on:
    • webcam (--source 0, 1, …)
    • image files
    • video files
    • generic video/stream URLs (RTSP/HTTP, etc.)

• 🧠 YOLOv8 integration

  • Uses Ultralytics YOLO (yolov8n.pt by default)
  • Supports custom .pt models via --model

• 📊 Visual overlays

  • Bounding boxes and class labels (from YOLO)
  • Per-class object counts in the top-left corner
  • FPS overlay (inference FPS) at the bottom of the frame

• 💾 Saving output (optional)

  • --save to write annotated images/videos
  • --out to control output path
  • --fps to control output video FPS (for streams/videos)

• 🖥️ CLI-focused design

  • --conf, --imgsz, --device, --show, --verbose for fine-grained control
  • start.bat for quick webcam testing on Windows

• ⚡ Fast startup

  • OpenCV, Ultralytics, and NumPy are imported lazily to avoid long initial load times.

---

📦 Requirements

• Python: 3.8 – 3.12

• Install dependencies:

  pip install -r requirements.txt
  

---

🚀 Quick start

Webcam (default camera 0):

python detect.py --source 0 --show

or on Windows:

start.bat

Image file:

python detect.py --source path/to/image.jpg --show --save

Video file:

python detect.py --source path/to/video.mp4 --show --save --fps 30

---

⚠️ Known limitations (v0.1)

This is a prototype, so expect rough edges:

• No tracking / IDs between frames yet (pure per-frame detection).
• No config file or presets; everything is CLI-driven.
• No automatic GPU/CPU diagnostics beyond basic logging.
• No tests, no packaging, just a single script.

---

If you try this and it breaks, that’s expected at v0.1 – Prototype.
Open an issue with your OS, Python version, command used, and console output.