Angular 6+ & OpenCV.js integration service library
Switch branches/tags
Nothing to show
Clone or download


This is a library that integrates Angular v6+ with OpenCVJS, the Javascript port of the popular computer vision library. It will allow you to load the library (with its WASM components) and use it in your application. The loading is done asynchrosnously after your Angular app has booted. The attached service makes use of a notifier to indicate when the loading is done and the service and library is ready for use.

Please read this blog post for the whole background on how this library came together



npm install ng-open-cv --save


yarn add ng-open-cv

Once the library is installed you will need to copy the opencv content from the node_modules/ng-open-cv/lib/assets folder to your own assets folder. This folder contains the actual OpenCV library (v3.4) and its WASM and ASM.js files.

Data files

OpenCV.js uses classification files to perform certain detection operations. To use those files:

  • Get the folders you need from the OpenCV data repository
  • Add the folders to your app's assets\opencv\data folder. Right now the assets\data folder that with this library only includes the haarcascardes files.
  • Use the createFileFromUrl in the NgOpenService class to load the file in memory.



  1. Typings

To your src/typings.d.ts file. add

declare var cv: any;


You can visit the demo site here.


If you have installed NgOpenCV and copied the opencv folder to your assets directory, everything should work out of the box.

1. Import the NgOpenCVModule

Create the configuration object needed to configure the loading of OpenCV.js for your application. By default the 3.4 asm.js version of the library will be loaded. The default options are

    scriptUrl: 'assets/opencv/asm/3.4/opencv.js',
    usingWasm: false,
    locateFile: this.locateFile.bind(this),
    onRuntimeInitialized: () => {}

Adjust the scriptUrl to contain the path to your opencvjs file. If you wanted to load the WASM version, you would use a configuration like:

const openCVConfig: OpenCVOptions = {
  scriptUrl: `assets/opencv/wasm/3.4/opencv.js`,
  wasmBinaryFile: 'wasm/3.4/opencv_js.wasm',
  usingWasm: true

Note: WASM is not supported on mobile Safari

Import NgOpenCVModule.forRoot(config) in the NgModule of your application. The forRoot method is a convention for modules that provide a singleton service. Pass it the configuration object.

import { BrowserModule } from '@angular/platform-browser';
import { NgModule } from '@angular/core';

import { AppComponent } from './app.component';
import { NgOpenCVModule } from 'ng-open-cv';
import { RouterModule } from '@angular/router';
import { AppRoutingModule } from './app-routing.module';
import { OpenCVOptions } from 'projects/ng-open-cv/src/public_api';

const openCVConfig: OpenCVOptions = {
  scriptUrl: `assets/opencv/opencv.js`,
  wasmBinaryFile: 'wasm/opencv_js.wasm',
  usingWasm: true

   declarations: [
   imports: [
   providers: [],
   bootstrap: [
export class AppModule { }

If you have multiple NgModules and you use one as a shared NgModule (that you import in all of your other NgModules), don't forget that you can use it to export the NgOpenCVModule that you imported in order to avoid having to import it multiple times.

const openCVConfig: OpenCVOptions = {
  scriptUrl: `assets/opencv/opencv.js`,
  wasmBinaryFile: 'wasm/opencv_js.wasm',
  usingWasm: true

    imports: [
    exports: [BrowserModule, NgOpenCvModule],
export class SharedModule {

3. Use the NgOpenCVService for your application

  • Import NgOpenCVService from ng-open-cv in your application code:
import { Component, OnInit, ElementRef, ViewChild } from '@angular/core';
import { fromEvent, Observable } from 'rxjs';
import { switchMap } from 'rxjs/operators';
import { NgOpenCVService, OpenCVLoadResult } from 'ng-open-cv';

  selector: 'app-hello',
  templateUrl: './hello.component.html',
  styleUrls: ['./hello.component.css']
export class HelloComponent implements OnInit {

  // Keep tracks of the ready
  openCVLoadResult: Observable<OpenCVLoadResult>;

  // HTML Element references
  fileInput: ElementRef;
  canvasOutput: ElementRef;

  constructor(private ngOpenCVService: NgOpenCVService) { }

  ngOnInit() {
    this.openCVLoadResult = this.ngOpenCVService.isReady$;

  loadImage(event) {
    if ( {
      const reader = new FileReader();
      const load$ = fromEvent(reader, 'load');
          switchMap(() => {
            return this.ngOpenCVService.loadImageToHTMLCanvas(`${reader.result}`, this.canvasOutput.nativeElement);
          () => {},
          err => {
            console.log('Error loading image', err);


The NgOpenCVService exposes a isReady$ observable which you should always subscribe too before attempting to do anything OpenCV related. It emits an OpenCVLoadResult object that is structured as:

export interface OpenCVLoadResult {
 ready: boolean;
 error: boolean;
 loading: boolean;

The following function gives you an example of how to use it your code:

detectFace() {
   // before detecting the face we need to make sure that
   // 1. OpenCV is loaded
   // 2. The classifiers have been loaded
       filter((result: OpenCVLoadResult) => result.ready),
       switchMap(() => {
         return this.classifiersLoaded$;
       tap(() => {
     .subscribe(() => {
       console.log('Face detected');

You can view more of this example code in the Face Detection Component


Run ng build to build the project. The build artifacts will be stored in the dist/ directory. Use the --prod flag for a production build.

Running unit tests

Run ng test to execute the unit tests via Karma.

Running end-to-end tests

Run ng e2e to execute the end-to-end tests via Protractor.



How to build a library for Angular apps