https://github.com/sammdu/pyimgscan

Take any phone-taken picture and turn it into a document scan.
https://github.com/sammdu/pyimgscan

image-processing opencv perspective-correction perspective-transform productivity python python3 scan scanner

Last synced: about 1 month ago
JSON representation

Take any phone-taken picture and turn it into a document scan.

• Host: GitHub
• URL: https://github.com/sammdu/pyimgscan
• Owner: sammdu
• License: mit
• Created: 2019-03-05T03:52:04.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2022-06-22T13:43:39.000Z (about 2 years ago)
• Last Synced: 2024-05-02T17:14:47.792Z (5 months ago)
• Topics: image-processing, opencv, perspective-correction, perspective-transform, productivity, python, python3, scan, scanner
• Language: Python
• Homepage: https://gitlab.com/sammdu/pyimgscan
• Size: 9.52 MB
• Stars: 75
• Watchers: 3
• Forks: 8
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# pyimgscan: open-source document scanning tool ![MIT License](https://img.shields.io/badge/LICENSE-MIT-brightgreen.svg?style=for-the-badge)

### Take a picture with your phone and convert it to a professional-looking scan.



### * Usage

**`$ python3.9 ./pyimgscan.py`**

```

usage: pyimgscan.py [-h] -i IMAGE [-I [INVERTED]]

```

The program will by default output a `corrected.png` and a `thresholded.png`

* `-h` display the above help message

* `-i` specify a path to the source image file

* `-I` also output an inverted binary image (white on black) to `thresholded_inverted.png`




### * Dependencies

![Python3.9](https://img.shields.io/badge/python-3.9-blue.svg?style=for-the-badge)

```

numpy==1.16.2

opencv-contrib-python==4.2.0.32

```

can also be found at `requirements.txt`




### * Setup with `virtualenv`

**1. Set up virtual environment:**   

```

virtualenv -p python3.9 PyImgScan/

cd ./PyImgScan/

./bin/activate

```

**2. Clone repository:**   

![GitLab](https://img.shields.io/badge/REPO-GitLab-blueviolet.svg?style=for-the-badge)   

`git clone https://gitlab.com/sammdu/pyimgscan.git`    

![GitHub](https://img.shields.io/badge/repo-github-lightgrey.svg?style=for-the-badge)   

`git clone https://github.com/sammdu/pyimgscan.git`

**3. Install dependencies:**   

```

cd pyimgscan/

pip install -r requirements.txt

```

**4. Test out an example!**   

```

python ./pyimgscan.py -i ./test/1.png

```

Available examples: `./test/1.png` `./test/2.jpg` `./test/3.jpg`




### * TODO

* unify average brightness for better adaptation to different lighting of input images

* instead of convex hull, try to always complete a 4-corner-polygon

* try Gaussian thresholding for better binary image results   

  https://docs.opencv.org/3.4.0/d7/d4d/tutorial_py_thresholding.html   

* add shadow removal functionality   

  https://stackoverflow.com/questions/44047819/increase-image-brightness-without-overflow/44054699#44054699   

  https://stackoverflow.com/questions/44752240/how-to-remove-shadow-from-scanned-images-using-opencv   

* add a border / four anchor points for better positioning and more reliable contour recognition

* build binaries with PyInstaller

* support multi-color thresholding




* * *

### References:

`imutils` https://github.com/jrosebr1/imutils

`PyImageSearch` https://www.pyimagesearch.com/2014/09/01/build-kick-ass-mobile-document-scanner-just-5-minutes/