Morph1x – YOLO Object Detection

A lightweight object detection system built on YOLOv8 with real-time tracking and video processing.

---

Features

• Real-time object detection with YOLOv8
• Persistent object tracking using ByteTrack
• Webcam and video file support
• Video output with multiple codec support
• Configurable confidence and IoU thresholds
• Multi-class detection with filtering
• CUDA GPU acceleration
• YAML-based configuration
• FPS monitoring

---

Requirements

• Python 3.8+
• CUDA-capable GPU (optional)

pip install -r requirements.txt

---

Installation

git clone https://github.com/yourusername/morph1x-yolo-detection.git
cd morph1x-yolo-detection
pip install -r requirements.txt

---

Usage

Webcam Detection

python run.py --source 0

Video File

python run.py --source "data/video.mp4"

Save Output

python run.py --source 0 --output "output/result.mp4"

Custom Settings

python run.py --source 0 --conf 0.6 --iou 0.5 --classes person car dog

Batch Process Video

python -m src.lib.process_video --input "data/input.mp4" --output "output/processed.mp4"

Press q to quit.

---

Arguments

run.py

• --source - Video source (0 for webcam or file path)
• --model - YOLO model path (default: model/yolov8n.pt)
• --classes - Allowed classes (default: person car dog)
• --output - Output video path
• --conf - Confidence threshold (default: 0.5)
• --iou - IoU threshold (default: 0.45)

process_video.py

• --input - Input video path (required)
• --output - Output video path (required)
• --model - YOLO model path
• --classes - Allowed classes
• --conf - Confidence threshold
• --iou - IoU threshold

---

Configuration

Edit config/model_config.yaml:

model:
  path: "models/current/yolov8n.pt"

inference:
  confidence_threshold: 0.5
  iou_threshold: 0.45
  max_detections: 300
  device: "cuda"  # or "cpu"

tracking:
  tracker: "bytetrack"
  max_age: 30
  min_hits: 3

visualization:
  colors:
    person: [0, 200, 0]
    car: [0, 180, 255]
    dog: [255, 120, 50]

---

Project Structure

morph1x-yolo-detection/
├── config/
│   └── model_config.yaml
├── src/
│   ├── core/
│   │   ├── object_detector.py
│   │   ├── Post_processing.py
│   │   └── renderer.py
│   └── lib/
│       └── process_video.py
├── model/
│   └── yolov8n.pt
├── run.py
└── requirements.txt

---

Supported Classes

Supports all 80 COCO classes: person, bicycle, car, motorcycle, airplane, bus, train, truck, boat, traffic light, fire hydrant, stop sign, parking meter, bench, bird, cat, dog, horse, sheep, cow, elephant, bear, zebra, giraffe, backpack, umbrella, handbag, tie, suitcase, frisbee, skis, snowboard, sports ball, kite, baseball bat, baseball glove, skateboard, surfboard, tennis racket, bottle, wine glass, cup, fork, knife, spoon, bowl, banana, apple, sandwich, orange, broccoli, carrot, hot dog, pizza, donut, cake, chair, couch, potted plant, bed, dining table, toilet, tv, laptop, mouse, remote, keyboard, cell phone, microwave, oven, toaster, sink, refrigerator, book, clock, vase, scissors, teddy bear, hair drier, toothbrush.

Filter with --classes argument.

---

Key Features Explained

Tracking

Objects maintain consistent IDs across frames using ByteTrack. Each detection includes track_id and unique_id (e.g., "ID-23-P").

Post-Processing

• Confidence-based filtering
• IoU-based Non-Maximum Suppression per class
• Maximum detection count limiting

Video Output

Automatic codec fallback: MP4V → H264 → MJPG → XVID → Default

---

Performance

Speed:

python run.py --source 0 --model model/yolov8n.pt --conf 0.6

Accuracy:

python run.py --source video.mp4 --model model/yolov8x.pt --conf 0.3

Check GPU: import torch; torch.cuda.is_available()

---

Troubleshooting

Model not found: Ensure model exists at model/yolov8n.pt

Can't open video: Check camera index (0, 1, 2) or file path

Low FPS: Use smaller model (yolov8n), lower confidence, enable GPU

Codec error: Try .avi extension or different output path