🏙️ SmartCity Vision Pipeline

A production-ready, modular Edge-AI video analytics pipeline designed for real-time vehicle detection, multi-object tracking, and region-crossing analytics. Built with Python and optimized for Video Management Software (VMS) ecosystems.

🚀 Key Features

• Multi-Class Vehicle Detection: Leverages an optimized YOLOv8 framework to detect and classify urban traffic entities (cars, trucks, buses, motorcycles) with high precision.
• Hardware-Aware Execution: Automatically detects and utilizes NVIDIA CUDA cores if available, falling back to CPU gracefully. Includes a built-in Frame Stride optimizer for low-end edge devices.
• Robust Multi-Object Tracking (MOT): Utilizes the ByteTrack algorithm to maintain consistent unique IDs across frames, handling occlusions and complex overlaps.
• Dynamic Aspect Ratio Handling: Flawlessly processes both traditional CCTV (16:9) and mobile/vertical (9:16) streams without distorting bounding box geometries.
• Privacy-by-Design (GDPR Compliant): Pre-architected with a Real-Time Anonymization Layer that blurs localized sensitive regions to conform strictly with the EU AI Act.

⚙️ Configuration Guide

The pipeline is designed to be plug-and-play. You can easily adjust the core mechanics by modifying the configuration block at the top of src/main.py:

# --- USER CONFIGURATION ---
VIDEO_SOURCE = 'data/traffic.mp4'  # Supports local paths or RTSP streams
MODEL_TYPE = 'yolov8s.pt'          # Choose 'n', 's', or 'm' based on your GPU
FRAME_STRIDE = 1                   # Increase to 2 or 3 for CPU optimization
CONFIDENCE = 0.15                  # Detection sensitivity threshold

🛠️ Tech Stack & Dependencies
Core Engine: Python 3.10+

Inference & Deep Learning: ultralytics, torch

Computer Vision: opencv-python, numpy

Tracking & Annotations: supervision

🚀 Installation & Quick Start
Clone the Repository:

Bash
git clone [https://github.com/yourusername/smartcity-vision-pipeline.git](https://github.com/yourusername/smartcity-vision-pipeline.git)
cd smartcity-vision-pipeline
Create a Virtual Environment (Recommended):

Bash
python -m venv venv
source venv/bin/activate  # On Windows use: venv\Scripts\activate
Install Dependencies:

Bash
pip install -r requirements.txt
Run the Pipeline:
Drop a test video named traffic.mp4 into the data/ folder, then run:

Bash
python src/main.py