Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/utkarsh-deshmukh/single-image-dehazing-python
python implementation of the paper: "Efficient Image Dehazing with Boundary Constraint and Contextual Regularization"
https://github.com/utkarsh-deshmukh/single-image-dehazing-python
airlight-estimation computer-vision defogging fog-removal haze-removal haze-removal-algorithm ieee image-dehazing opencv python single-image-defogging single-image-dehazing
Last synced: 27 days ago
JSON representation
python implementation of the paper: "Efficient Image Dehazing with Boundary Constraint and Contextual Regularization"
- Host: GitHub
- URL: https://github.com/utkarsh-deshmukh/single-image-dehazing-python
- Owner: Utkarsh-Deshmukh
- License: bsd-2-clause
- Created: 2020-06-24T06:06:06.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2024-05-30T02:15:46.000Z (5 months ago)
- Last Synced: 2024-06-21T15:53:59.248Z (5 months ago)
- Topics: airlight-estimation, computer-vision, defogging, fog-removal, haze-removal, haze-removal-algorithm, ieee, image-dehazing, opencv, python, single-image-defogging, single-image-dehazing
- Language: Python
- Homepage:
- Size: 7.52 MB
- Stars: 95
- Watchers: 3
- Forks: 24
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Single-Image-Dehazing-Python
python implementation of the paper: "Efficient Image Dehazing with Boundary Constraint and Contextual Regularization"
## Installation and Running the tests
### method 1
```
pip install image_dehazer
```
**Usage:**
```
import image_dehazer # Load the library
HazeImg = cv2.imread('image_path') # read input image -- (**must be a color image**)
HazeCorrectedImg, HazeTransmissionMap = image_dehazer.remove_haze(HazeImg) # Remove Haze
cv2.imshow('input image', HazeImg); # display the original hazy image
cv2.imshow('enhanced_image', HazeCorrectedImg); # display the result
cv2.waitKey(0) # hold the display window
```
### user controllable parameters (with their default values):
```
airlightEstimation_windowSze=15
boundaryConstraint_windowSze=3
C0=20
C1=300
regularize_lambda=0.1
sigma=0.5
delta=0.85
showHazeTrasmissionMap=True
```
### method 2
1. Go to the src folder
2. run the file "example.py"
3. sample images are stored in the "Images/" folder
4. Output images will be stored in the "outputImages/" folder
# Libraries needed:
1.numpy==1.19.0
2.opencv-python
3.scipy
# Theory
This code is an implementation of the paper "Efficient Image Dehazing with Boundary Constraint and Contextual Regularization"
The algorithm can be divided into 4 parts:
- Airlight estimation
- Calculating boundary constraints
- Estimate and refine Transmission
- Perform Dehazing using the estimated Airlight and Transmission
# License
- This project is licensed under the BSD 2 License - see the LICENSE.md file for details
# Acknowledgements
- The author would like to thank "Gaofeng MENG" and his implementation of his algorithm: https://github.com/gfmeng/imagedehaze
- The author would like to thank Gaofeng MENG, Ying WANG, Jiangyong DUAN, Shiming XIANG, Chunhong PAN for their paper "Efficient Image Dehazing with Boundary Constraint and Contextual Regularization"
- The author would like to thank Alexandre Boucaud. The function psf2otf was obtained from his repository. (https://github.com/aboucaud/pypher/blob/master/pypher/pypher.py)
- The Author would like to thank Dr. Suresh Merugu for his matlab implementation of the codes. This repository is the python implementation of the matlab codes.
- The Author would like to thank Mayank Singal for his repository "[PyTorch-Image-Dehazing](https://github.com/MayankSingal/PyTorch-Image-Dehazing)" which gives a pytorch implementation of the AOD-Net architecture. [Link to ICCV 2017 paper](https://openaccess.thecvf.com/content_ICCV_2017/papers/Li_AOD-Net_All-In-One_Dehazing_ICCV_2017_paper.pdf)
Merugu, Suresh. (2014). Re: How to detect fog in an image and then enhance the image to remove fog?. Retrieved from: https://www.researchgate.net/post/How_to_detect_fog_in_an_image_and_then_enhance_the_image_to_remove_fog/53ae3f10d2fd64c3648b45a9/citation/download.
# Citation
```
@INPROCEEDINGS{6751186,
author={G. Meng and Y. Wang and J. Duan and S. Xiang and C. Pan},
booktitle={IEEE International Conference on Computer Vision},
title={Efficient Image Dehazing with Boundary Constraint and Contextual Regularization},
year={2013},
volume={},
number={},
pages={617-624},
month={Dec},}
```
# Results
============================================================================================================
============================================================================================================
# Performance Comparison:
In this section, I am comparing the dehazing output with that of AOD-Net. I am using this [python implementation of AOD-Net](https://github.com/MayankSingal/PyTorch-Image-Dehazing/tree/master) to run a pretrained AOD-Net model
Here are some cases where AOD-Net is better:
