OpenCV-Wasm

Precompiled OpenCV to JavaScript + WebAssembly for node.js and deno environment. 🦕

In this Wasm-compiled OpenCV, there’s no need to have OpenCV installed in the machine. The entire OpenCV library is already inside this package (opencv.js and opencv.wasm).

This module has zero dependencies.

Examples

Code	Input	Output
dilation.js (node)
templateMatching.js (node)	source: template:

Installation

node

npm install opencv-wasm

Code example:

const { cv, cvTranslateError } = require('./index.js');

let mat = cv.matFromArray(2, 3, cv.CV_8UC1, [1, 2, 3, 4, 5, 6]);
console.log('cols =', mat.cols, '; rows =', mat.rows);
console.log(mat.data8S);

cv.transpose(mat, mat);
console.log('cols =', mat.cols, '; rows =', mat.rows);
console.log(mat.data8S);

/*
cols = 3 ; rows = 2
Int8Array(6) [ 1, 2, 3, 4, 5, 6 ]
cols = 2 ; rows = 3
Int8Array(6) [ 1, 4, 2, 5, 3, 6 ]
*/

deno

import { cv, cvTranslateError } from 'https://deno.land/x/opencv@v4.3.0-10/mod.ts';
// Change the @<version> with the latest or any version you desire.
// Check the available versions here: https://deno.land/x/opencv.

Code example:

import { cv, cvTranslateError } from 'https://deno.land/x/opencv@v4.3.0-10/mod.ts';

let mat = cv.matFromArray(2, 3, cv.CV_8UC1, [1, 2, 3, 4, 5, 6]);
console.log('cols =', mat.cols, '; rows =', mat.rows);
console.log(mat.data8S);

cv.transpose(mat, mat);
console.log('cols =', mat.cols, '; rows =', mat.rows);
console.log(mat.data8S);

/*
cols = 3 ; rows = 2
Int8Array(6) [ 1, 2, 3, 4, 5, 6 ]
cols = 2 ; rows = 3
Int8Array(6) [ 1, 4, 2, 5, 3, 6 ]
*/

Usage

Because this module is using the same code as the official OpenCV.js for the web, you can use the same documentation at the web: https://docs.opencv.org/4.3.0/d5/d10/tutorial_js_root.html

There are some minor initialization changes, because this module will be loaded synchronously instead of the OpenCV’s default (asynchronously).

You can check the files inside examples folder as reference on how to initialize, loading images, and saving images.

Error Handling

By default, mistakes in code will produce error code. You can use the following snippet to translate the error code into meaningful statement from OpenCV.

const { cv, cvTranslateError } = require('opencv-wasm');

try {
    // Your OpenCV code
} catch (err) {
    console.log(cvTranslateError(cv, err));
}

Versioning

This npm module uses the following versioning number:

<opencv version>-<this module version>

For Example

4.3.0-9
OpenCV version 4.3.0
OpenCV-Wasm Module version 9

Development

Building

Run the following script on macOS or Linux (tested on Ubuntu). You need docker on the system.

npm install
(cd ./utils && sh ./build.sh)
(cd utils && node generateCvProps.js)

Testing

After completing the build script, you can run the test provided by OpenCV, and the test from this repo.

# OpenCV's test
(cd ./build_wasm_test/bin && npm install)
(cd ./build_wasm_test/bin && node tests.js)

# This repo's test
npm test

Authors

Ezzat Chamudi - echamudi

See also the list of contributors who participated in this project.

License

OpenCV-Wasm code is licensed under BSD-3-Clause. Images, logos, docs, and articles in this project are released under CC-BY-SA-4.0.

Libraries, dependencies, and tools used in this project are tied with their licenses.