Emscripten port of ARToolKit to JavaScript.
JSARToolKit5 support these types of markers:
- Square pictorial markers
- Square barcode markers
- Multi square markers set
- NFT (natural feature tracking) markers
NOTE:
When writing JavaScript and making changes be aware that the emscripten uglifier does not support the ES6 syntax.
build/
(compiled debug and minified versions of ARToolKit.js)doc/
(documentation, coming...)emscripten/
(source code for ARToolKit.js)examples/
(demos and examples using ARToolKit.js)js/
(compiled versions of ARToolKit.js with Three.js helper api)tools/
(build scripts for building ARToolKit.js)
JSARToolKit5 supports WebAssembly. The libary builds two WebAssembly artifacts during the build process. These are build/artoolkit_wasm.js
and build/artoolkit_wasm.wasm
. To use those, include the artoolkit_wasm.js into your html page and define var artoolkit_wasm_url = '<<PATH TO>>/artoolkit_wasm.wasm';
before loading the artoolkit_wasm.js file, like this:
<script type='text/javascript'>
var artoolkit_wasm_url = '../build/artoolkit_wasm.wasm';
</script>
<script src="../build/artoolkit_wasm.js"></script>
As loading the WebAssembly artifact is done asynchronously, there is a callback that is called when everything is ready.
window.addEventListener('artoolkit-loaded', () => {
//do artoolkit stuff here
});
See examples/simple_image_wasm.html for details.
- Clone this repository
- Clone ARToolKit5 project to get the latest source files. From within jsartoolkit5 directory do
git submodule update --init
. If you already cloned ARToolKit5 to a different directory you can:
- create a link in the
jsartoolkit5/emscripten/
directory that points to ARToolKit5 (jsartoolkit5/emscripten/artoolkit5
) (Linux and macOS only) - or, set the
ARTOOLKIT5_ROOT
environment variable to point to your ARToolKit5 clone - or, change the
tools/makem.js
file to point to your artoolkit5 clone (line 20)
- Install Docker (if you havn't already): get Docker
- Clone artoolkit5 repository on your machine:
git submodule update --init
npm install
- From inside jsartoolkit5 directory run
docker run -dit --name emscripten -v $(pwd):/src trzeci/emscripten-slim:latest bash
to download and start the container, in preparation for the build docker exec emscripten npm run build-local
to build JS version of artoolkit5docker stop emscripten
to stop the container after the build, if neededdocker rm emscripten
to remove the containerdocker rmi trzeci/emscripten-slim:latest
to remove the Docker image, if you don't need it anymore- The build artifacts will appear in
/build
. There's a build with debug symbols inartoolkit.debug.js
file and the optimized build with bundled JS API inartoolkit.min.js
; also, a WebAssembly build artoolkit_wasm.js and artoolkit_wasm.wasm
To prevent issues with Emscripten setup and to not have to maintain several build environments (macOS, Windows, Linux) we only maintain the Build using Docker. Following are the instructions of the last know build on Linux which we verified are working. Use at own risk. ** Not working on macOS!**
- Install build tools
- Install node.js (https://nodejs.org/en/)
- Install python2 (https://www.python.org/downloads/)
- Install emscripten (https://emscripten.org/docs/getting_started/downloads.html#download-and-install)
We used emscripten version 1.39.5-fastcomp
1.38.44-fastcomp
jsartoolkit5 aim is to create a Javascript version of artoolkit5. First, you need the artoolkit5 repository on your machine:
2. Clone ARToolKit5 project to get the latest source files. From within jsartoolkit5 directory do git submodule update --init
. If you already cloned ARToolKit5 to a different directory you can:
- create a link in the
jsartoolkit5/emscripten/
directory that points to ARToolKit5 (jsartoolkit5/emscripten/artoolkit5
) - or, set the
ARTOOLKIT5_ROOT
environment variable to point to your ARToolKit5 clone - or, change the
tools/makem.js
file to point to your artoolkit5 clone (line 20)
- Building
- Make sure
EMSCRIPTEN
env variable is set (e.g.EMSCRIPTEN=/usr/lib/emsdk_portable/emscripten/master/ node tools/makem.js
- Run
npm install
- Run
npm run build-local
During development, you can run npm run watch
, it will rebuild the library everytime you change ./js/
directory.
- The built ASM.js files are in
/build
. There's a build with debug symbols inartoolkit.debug.js
and the optimized build with bundled JS API inartoolkit.min.js
.
<script src="../build/artoolkit.min.js">
// include optimized ASM.js build and JS API
</script>
<script async src="../build/artoolkit.debug.js">
// - include debug build
</script>
<script src="../js/artoolkit.api.js">
// - include JS API
</script>
<script src="../build/artoolkit.min.js">
// - include optimized ASM.js build and JS API
</script>
<script src="js/third_party/three.js/three.min.js">
// - include Three.js
</script>
<script src="../js/artoolkit.three.js">
// - include Three.js helper API
</script>
<script>
window.ARThreeOnLoad = function () {
console.log("Three.js helper API loaded");
};
if (window.ARController && window.ARController.getUserMediaThreeScene) {
ARThreeOnLoad();
};
</script>
See examples/
for examples on using the raw API and the Three.js helper API.
The basic operation goes like this:
- Load a
ARCameraParam
object - Create a
ARController
object - Set pattern detection mode
- Load pattern markers or multimarkers if needed
- Add a
'getMarker'
event listener - Call
ARController.process(img)
<script src="../build/artoolkit.min.js"></script>
<script>
var param = new ARCameraParam();
param.onload = function () {
var img = document.getElementById('my-image');
var ar = new ARController(img.width, img.height, param);
// Set pattern detection mode to detect both pattern markers and barcode markers.
// This is more error-prone than detecting only pattern markers (default) or only barcode markers.
//
// For barcode markers, use artoolkit.AR_MATRIX_CODE_DETECTION
// For pattern markers, use artoolkit.AR_TEMPLATE_MATCHING_COLOR
//
ar.setPatternDetectionMode(artoolkit.AR_TEMPLATE_MATCHING_COLOR_AND_MATRIX);
ar.addEventListener('markerNum', function (ev) {
console.log('got markers', markerNum);
});
ar.addEventListener('getMarker', function (ev) {
console.log('found marker?', ev);
});
ar.loadMarker('Data/patt.hiro', function (marker) {
console.log('loaded marker', marker);
ar.process(img);
});
};
param.src = 'Data/camera_para.dat';
</script>
NFT (Natural Feature Tracking) is a markerless technology that let you track almost any images you want. To use this feature take a look at the nft_improved_worker example folder. If you want to create your custom NFT marker you can use the online tool NFT-Marker-Creator. Before proceeding with the creation of your markers, carefully read the information on the wiki.
In the code below a summarized example:
<div id="container">
<video id="video"></video>
<canvas style="position: absolute; left:0; top:0" id="canvas_draw"></canvas>
</div>
// main worker create the web worker see in the examples/nft_improved_worker for details
<script src="main_worker.js"></script>
<script>
var container = document.getElementById('container');
var video = document.getElementById('video');
var canvas_draw = document.getElementById('canvas_draw');
if (navigator.mediaDevices && navigator.mediaDevices.getUserMedia) {
var hint = {};
if (isMobile()) {
hint = {
facingMode: {"ideal": "environment"},
audio: false,
video: {
width: {min: 240, max: 240},
height: {min: 360, max: 360},
},
};
}
navigator.mediaDevices.getUserMedia({video: hint}).then(function (stream) {
video.srcObject = stream;
video.play();
video.addEventListener("loadedmetadata", function() {
start(container, markers["pinball"], video, video.videoWidth, video.videoHeight, canvas_draw, function() { statsMain.update() }, function() { statsWorker.update()) };
});
});
}
</script>
prepend all these constants with Module.
or artoolkit.CONSTANTS
to access them
- AR_DEBUG_DISABLE
- AR_DEBUG_ENABLE
- AR_DEFAULT_DEBUG_MODE
- AR_LABELING_WHITE_REGION
- AR_LABELING_BLACK_REGION
- AR_DEFAULT_LABELING_MODE
- AR_DEFAULT_LABELING_THRESH
- AR_IMAGE_PROC_FRAME_IMAGE
- AR_IMAGE_PROC_FIELD_IMAGE
- AR_DEFAULT_IMAGE_PROC_MODE
- AR_TEMPLATE_MATCHING_COLOR
- AR_TEMPLATE_MATCHING_MONO
- AR_MATRIX_CODE_DETECTION
- AR_TEMPLATE_MATCHING_COLOR_AND_MATRIX
- AR_TEMPLATE_MATCHING_MONO_AND_MATRIX
- AR_DEFAULT_PATTERN_DETECTION_MODE
- AR_USE_TRACKING_HISTORY
- AR_NOUSE_TRACKING_HISTORY
- AR_USE_TRACKING_HISTORY_V2
- AR_DEFAULT_MARKER_EXTRACTION_MODE
- AR_MAX_LOOP_COUNT
- AR_LOOP_BREAK_THRESH
- AR_MATRIX_CODE_3x3
- AR_MATRIX_CODE_3x3_HAMMING63 5
- AR_MATRIX_CODE_3x3_PARITY65 2
- AR_MATRIX_CODE_4x4
- AR_MATRIX_CODE_4x4_BCH_13_9_3 7
- AR_MATRIX_CODE_4x4_BCH_13_5_5 10
- AR_LABELING_THRESH_MODE_MANUAL
- AR_LABELING_THRESH_MODE_AUTO_MEDIAN
- AR_LABELING_THRESH_MODE_AUTO_OTSU
- AR_LABELING_THRESH_MODE_AUTO_ADAPTIVE
- AR_MARKER_INFO_CUTOFF_PHASE_NONE
- AR_MARKER_INFO_CUTOFF_PHASE_PATTERN_EXTRACTION
- AR_MARKER_INFO_CUTOFF_PHASE_MATCH_GENERIC
- AR_MARKER_INFO_CUTOFF_PHASE_MATCH_CONTRAST
- AR_MARKER_INFO_CUTOFF_PHASE_MATCH_BARCODE_NOT_FOUND
- AR_MARKER_INFO_CUTOFF_PHASE_MATCH_BARCODE_EDC_FAIL
- AR_MARKER_INFO_CUTOFF_PHASE_MATCH_CONFIDENCE
- AR_MARKER_INFO_CUTOFF_PHASE_POSE_ERROR
- AR_MARKER_INFO_CUTOFF_PHASE_POSE_ERROR_MULTI
- AR_MARKER_INFO_CUTOFF_PHASE_HEURISTIC_TROUBLESOME_MATRIX_CODES
You can run an automated routine to make some tests, in the main jsartoolkit5 folder just run in a console the command:
npm run test
Then open the tests page:
http://localhost:8085/tests/index.html
It is possible to build the api documentation, run this command in the main jsartoolkit5 folder:
npm run create-doc
The api documentation will be created in the doc folder. Navigate to the reference folder to view the api docs.
If you found a bug or you have a feature request, or for any inquiries related to jsartoolkit5 development file an issue at: