Skip to content

Latest commit

 

History

History
150 lines (124 loc) · 3.88 KB

local-setup.md

File metadata and controls

150 lines (124 loc) · 3.88 KB

🎯 Library Seat Detection - Complete Setup Guide

📋 Prerequisites

  • Python 3.8+ installed
  • ~3GB free disk space
  • Internet connection for model downloads

🚀 Setup Instructions

1. Navigate to project

cd /home/priyanshu/repos/Library-Seat-Occupancy-Detection

2. Create and activate virtual environment

# Create virtual environment (one-time setup)
python3 -m venv venv

# Activate virtual environment (do this every time)
source venv/bin/activate

3. Install dependencies

# Install all required packages (~3GB download)
pip install -r requirements.txt

⏳ This takes 10-15 minutes depending on internet speed

4. Verify installation

# Check if key packages are installed
python -c "import cv2, torch, numpy; print('✅ All packages installed successfully!')"

5. Run detection on your video

# Basic usage
python detect_and_track.py --source "library-demo-video.mp4"

# Advanced usage with custom settings
python detect_and_track.py \
    --weights yolov7.pt \
    --source "library-demo-video.mp4" \
    --conf-thres 0.4 \
    --classes 0 56 \
    --name "Library_Seat_Detection_Test"

6. Find your results

# Output video will be saved to:
ls runs/detect/*/

🔧 Command Line Arguments

Argument Description Default
--source Input video file Required
--weights Model weights file yolov7.pt
--conf-thres Confidence threshold 0.25
--classes Detect specific classes All classes
--name Output folder name object_tracking
--view-img Display real-time preview False

📊 Detected Classes

  • Class 0: Person (human detection)
  • Class 56: Chair (furniture detection)
  • Use --classes 0 56 to detect only people and chairs

🎯 Expected Output

  • Input: library-demo-video.mp4
  • Output: runs/detect/{experiment_name}/library-demo-video.mp4
  • Features:
    • Red boxes around people with tracking IDs
    • Yellow boxes around chairs
    • Occupancy time tracking
    • "TIME EXCEEDED" alerts for seats held too long

🛠️ Troubleshooting

ModuleNotFoundError: No module named 'cv2'

# Make sure virtual environment is activated
source venv/bin/activate

# Verify installation completed
pip list | grep opencv

Model weights not found

# First run will auto-download yolov7.pt (~74MB)
# Or download manually:
wget https://github.com/WongKinYiu/yolov7/releases/download/v0.1/yolov7.pt

Virtual environment issues

# Recreate virtual environment
rm -rf venv
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

📁 Project Structure

Library-Seat-Occupancy-Detection/
├── detect_and_track.py     # Main detection script
├── sort.py                 # Object tracking algorithm  
├── models/                 # YOLOv7 architecture
├── utils/                  # Helper functions
├── data/                   # Configuration files
├── requirements.txt        # Python dependencies
├── venv/                   # Virtual environment (created)
├── library-demo-video.mp4  # Your test video
└── runs/detect/            # Output videos (created)

🔄 Daily Workflow

# 1. Navigate and activate
cd /home/priyanshu/repos/Library-Seat-Occupancy-Detection
source venv/bin/activate

# 2. Run detection
python detect_and_track.py --source "your_video.mp4"

# 3. Check results
ls runs/detect/*/

# 4. Deactivate when done
deactivate

💾 Disk Space Usage

  • Project files: ~7MB
  • Virtual environment: ~2.2GB
  • YOLOv7 weights: ~74MB
  • Total: ~2.3GB

⚡ Performance Tips

  • Use --conf-thres 0.4 for better accuracy
  • Add --view-img to see real-time detection
  • Use smaller videos for testing (under 1GB)
  • Close other applications for better performance deactivate