Underwater image color correction and adaptive contrast enhancemention and yolo detect python

The original uses matlab and it p-code.

I used python to restore more than 80% of the algorithms in the paper.

And i add a method to exchange wrong channels problem.

For contrast enhancement, this repo have fusion method that can replace LACE.

This repo have webpage for apply MLLE and fusion. Webpage can upload image and download image after process.

This repo Webpage: MLLE-webpage

• Paper: Underwater Image Enhancement via Minimal Color Loss and Locally Adaptive Contrast Enhancement.

• Paper: Color Balance and Fusion for Underwater Image Enhancement

• Source matlab code: MLLE

• Test image from: UIEB dataset

Preview

Preview Video

Requirements

pip install -r requirements.txt

If you want to use gpu for object detection, please install gpu version pytorch.

How to use

Put the image(.jpg .png) or .mp4 in the Input folder.

This repo provides models for detecting sea turtles and sea urchins.

If you want to change the model please open the main.py file.

model = YOLO('to/your/model/path')

Usage example

Using the command line interface (CLI)

Choose mode MLLE or LACC+fusion

python main.py --mode mlle

or

python main.py -m fusion

Change to process video, detect object by yolo, beta

python main.py --type video --mode mlle --detect --beta 2

or

python main.py -t vid -m mlle -d -b 2