🚗 Python Yolov8 - Advanced Vehicle Counter

Christian Zigah — Fri, 04 Jul 2025 10:44:31 +0000

A Python-based advanced vehicle counting system using YOLOv8 and OpenCV. Tracks and counts vehicles crossing a virtual line, records output video, and logs detailed count statistics for each class in both directions.

[🎥 Watch Advance Car Counter - Output Recording on YouTube]

https://youtu.be/ha5XGrtyo9c

✨ Features

✅ YOLO-based object detection

✅ Class-specific counting: cars, trucks, buses, motorcycles

✅ Two-way counting: UP and DOWN directions

✅ Tracks individual vehicles by ID (without displaying ID on screen)

✅ Overlays detection boxes with class & confidence

✅ Video recording of processed output

✅ Detailed text logging with timestamps

✅ Real-time counting statistics panel

✅ Visual indicator for each crossing event

✅ Saves final summary report

🚀 Quick Start

Install dependencies:

pip install opencv-python ultralytics numpy

Download a YOLOv8 model (if you don't already have one):
- The default uses yolov8n.pt, available from Ultralytics YOLOv8 models.
Run the counter:
```
python main.py
```
By default, it processes traffic.mp4, drawing a counting line at 70% height.

🛠️ Usage

Create an instance of AdvancedVehicleCounter with your desired parameters:

from your_module import AdvancedVehicleCounter

counter = AdvancedVehicleCounter(
    model_path="yolov8n.pt",             # Path to your YOLOv8 model
    video_source="traffic.mp4",          # Path to your input video
    line_position=0.7,                   # Vertical position of the counting line (0-1 scale)
    output_folder="monitoring_results"   # Folder to save logs and processed video
)
counter.run()

📂 Outputs

Processed Video: Saved in the output folder with a timestamped filename (e.g., output_20250630_123456.mp4).
python-replCopyEditTimestamp,Class,Direction,TotalCount2025-06-30 12:34:56.789,car,DOWN,1...
python-replCopyEditFINAL SUMMARYClass,DOWN,UP,TOTALcar,10,8,18motorcycle,3,2,5...

📈 How It Works

🔹 DetectionUses YOLOv8 to detect vehicles in each frame.

🔹 TrackingMaintains a per-object path using YOLO’s tracker, enabling detection of line crossings.

🔹 CountingWhen a vehicle crosses the counting line:

UP: moves from below the line to above.
DOWN: moves from above the line to below.Counts are updated for the specific class.

🔹 Visualization

Draws bounding boxes with class name and confidence.
Highlights vehicles that crossed the line.
Displays a real-time panel with UP/DOWN counts for each class.

🔹 Output

Saves an annotated video to the output folder.
Logs every crossing event with timestamp, class, direction, and total count.
Appends a final summary at the end.

⚙️ Configuration Options

ParameterDescriptionDefaultmodel_pathYOLOv8 model weights file path"yolov8n.pt"video_sourceInput video file path"video2.mp4"line_positionVertical position of counting line (0-1 scale)0.5line_thicknessThickness of counting line in pixels2output_folderDirectory to save video and logs"output"

📌 Notes

Requires a GPU or fast CPU for real-time performance.
Tested with YOLOv8 and OpenCV 4.x.
Supports Linux, macOS, and Windows.
Press Q to exit the application early.

📸 Example

Once running, you will see:

Green detection boxes labeled with class & confidence.
A horizontal counting line at the configured position.
Visual indicators when a crossing happens (colored circles + text).
A panel showing real-time UP/DOWN counts per class.

Python Advance Car Counter on GitHub - Give me Stars

DEV Community: Christian Zigah