Haze Removal

Single Image Haze Removal Using Dark Channel Prior

Usages

启动服务

$ python server.py

使用CLI

# 恢复图片
$ python main.py recover --image images/gugong.bmp
# 恢复并保存图片
$ python main.py recover --image images/gugong.bmp --save
# 清除用户上传图片
$ python main.py clean
# benchmark
$ python main.py benchmark --tries 10 --image images/forest1.jpg

Docker Deploy

$ curl -O http://7vzsal.com1.z0.glb.clouddn.com/dehaze.tar.gz
$ tar -zxvf dehaze.tar.gz
$ docker build -t dehaze-server .
$ docker run -it --rm -p 5000:5000 --name my-dehaze-server dehaze-server

Install

$ pip install -r requirements.txt
$ python server.py

展示平台要求

运行中间结果　＋　除雾后的图