image-to-scan 0.0.4

Convert photos of documents made with a camera to a 'scanned' documents. It will take documents' contour and apply a four point transformation

Four-Point-Invoice-Transform-with-OpenCV

forked from KMKnation/Four-Point-Invoice-Transform-with-OpenCV

This code is inspired from [4 Point OpenCV getPerspective Transform Example]

I have customized the code of Adrian to find 4 points of document or rectangle dynamically. Here i have added findLargestCountours and convert_object, where convert_object is our driver method which actually doing image processing and getting all 4 point rectangles from image. After getting all 4 point rectangle list findLargestCountours method finding largest countour in list.

Installation

pip install image_to_scan

Run it

Before running the examples, create a virtual environment and install dependencies with make init, this will also add an entry point image-to-scan from which you can call the script.

Activate your virtualenv source venv/bin/activate.

Sample2

Run image-to-scan tests/samples/02/original.png

Original Image	Edge Detection	Warped Image

Sample3

Run image-to-scan tests/samples/03/original.png

Original Image	Edge Detection	Warped Image