Skip to content

Precompiled OpenCV 4.5.5 to JavaScript + WebAssembly for node and deno. 🦕

License

Notifications You must be signed in to change notification settings

CezaryDanielNowak/fast-opencv-wasm

 
 

Repository files navigation

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) image sample 1 dilation
templateMatching.js (node) source:
image sample 2
template:
image sample 2 template
template matching

Installation

node

npm install opencv-wasm

Code example:

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

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 ]
*/

Using in the browser

<script src="/opencv.js"></script>

<script>
function getBinaryPromise(wasmBinary) {
  return fetch(wasmBinary, {
    credentials: "same-origin",
    crossOrigin: "anonymous",
  }).then(function(response) {
    if (!response["ok"]) {
      throw "failed to load wasm binary file at '" + wasmBinary + "'"
    }
    return response["arrayBuffer"]();
  });
}

getBinaryPromise('./opencv.wasm').then((wasmBinary) => {
  const Module = {
    wasmBinary,
  };

  cv = cv(Module);
});
</script> 

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/4.5.5/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('fast-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

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

License

Copyright © 2020 Ezzat Chamudi and OpenCV-Wasm Project Authors

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.

OpenCV License.

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

About

Precompiled OpenCV 4.5.5 to JavaScript + WebAssembly for node and deno. 🦕

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 99.3%
  • Other 0.7%