Document Scanner

Scan an image for a document outline, crop it, and translate it into a rectangle.

Usage

The following assumes OpenCV.js is loaded and running in your environment and "mat" is a cv.Mat instance.

// create a DocumentScanner instance and provide the OpenCV object
const scanner = new DocumentScanner(cv);

// scan an OpenCV Mat instance for a document
// returns a Mat of the cropped and translated document
const doc = scanner.scanImage(
    // source Mat
    mat,
    // whether an outline of the document should be drawn onto the source Mat
    true
);

Example

Included in this repository is an example implemented with NodeJS which can be run within the /example directory via:

node index.js

By default a sample document is pre-loaded and scanned. You may upload your own images to test. Be sure that the image you upload had good contrast between the document and the background.

Thanks

• OpenCV for creating an amazing toolkit.
• PyImageSearch for a well documented and helpful tutorial to get things moving in the right direction.