Low-Light Object Detection & Enhancement

This project combines low-light image enhancement techniques with object detection using YOLOv8. It is designed to improve visibility in dark images before performing object detection.

Project Structure

• enhance.py: Core utility functions for image processing. Contains algorithms to artificially darken images and enhance low-light images using CLAHE (Contrast Limited Adaptive Histogram Equalization).
• make_dark.py: A demonstration script that takes an input image (bus.jpg), creates a darkened version (bus_dark.jpg), and then applies enhancement (bus_enhanced.jpg).
• train.py: Script to train a YOLOv8n model on a dataset (currently configured for coco128.yaml).
• configs/: Configuration files.
• inference.py: (Placeholder) For running inference on images/videos.
• evaluate.py: (Placeholder) For evaluating model performance.

Installation

1. Clone the repository.

git clone https://github.com/witharyank/Lowlight
cd lowlight

2. Install the required dependencies:

pip install -r requirements.txt

Usage

Image Enhancement Demo

To test the image enhancement capabilities:

python make_dark.py

This will generate:

• bus_dark.jpg: Artificially darkened image.
• bus_enhanced.jpg: The restored version of the dark image.

Training the Model

To start training the YOLOv8 model:

python train.py

Note: Ensure you have the necessary datasets configured in your YAML files.

Requirements

• Python 3.8+
• OpenCV
• Ultralytics YOLO
• NumPy
• (See requirements.txt for full list)

Author

Kumar Aryan GitHub: https://github.com/witharyank