Download

Augmented Reality for UE4

Community Work in Progress
#1

Plugin available on the link below:
http://ar.uplugins.com

[UPDATE]
AR version of the Matinee fight scene demo.:

Interaction DEMO video:

Hi everyone,

After working on this for quite some time, I wanted to show my first results.
I am developing a C++ code plugin that adds AR capabilities to Unreal Engine 4.

The videos below are featuring simple marker tracking examples based on the open source ARToolkit library. The DEMO can detect and track 2 markers at the same time.
Currently it is just a “proof of concept” running on Windows and Android, but I hope to release it as a plugin soon!

Inside UE4 editor on Windows:

Running on Samsung S4 mini:

Hope you like it!

#2

You sir, are a legend :slight_smile:

Looking forward to seeing how it goes…

#3

Yes yes yes yes so muuuch yessss!

#4

Awesome, You Rock!

#5

That is pretty big time impressive.

#6

OMG, I will buy this so hard!

#7

This is too good. Got to try this.

#8

Yeah, this is ridiculously awesome! Can’t wait to see more.

#9

Looks very promising, keep up the good work!

#10

Yeah, this is really amazing ! Keep it up !

#11

Wow,…loved it!

-m

#12

Will definitely be keeping up on this thread :slight_smile:

#13

impressive !!! when does your plugin will be in test or released ? and at wich price ? could be very usefull for my work !! thanks

#14

I’m starting to see some light at the end of the tunnel :slight_smile:
Keep it up!

#15

Hi,

I’ve uploaded a very early binary version of the upcoming AR plugin with a sample project.
Currently it works only in the UE editor (on Windows), and all of it’s features can be used with blueprints. No C++ coding required!

I’ve tried to keep everything as simple as possible.

Opinions / suggestion are welcomed!

How to test ?
===========

  1. Download the file on the link below and unzip it’s contents to your UE4 project directory.
    http://get.fitnect.hu/download.php?f=tools/ARSample.zip

  2. Print the markers located in ARSample\Plugins\ARToolkitPlugin\Binaries\Win64\Data\pdf

  • There are 4 markers here: Kanji , Hiro, Sample 1, Sample 2. (Print at least Hiro, to get started!)
  • Make sure to print them in their original size, and there should be a white border around the marker.

  1. Run Epic Games launcher, and install Unreal Engine 4.9.2

  2. In Epic Games launcher open ARSample project

  3. Connect your webcam, and press the play button inside the editor!

  4. Video stream should appear, move the markers (Hiro / Kanji) in front of the camera

  5. Default unreal 3D chairs should appear on the screen.

More info about marker tracking can be found on the following link:
http://artoolkit.org/documentation/doku.php?id=3_Marker_Training:marker_about

There are two main blueprints in the example:

ARToolkitBase blueprint
===================

  • Default webcam init (640x480 default resolution)
  • Initializes the video stream and creates a dynamic texture based on the baseFeed material
  • Loading markers (Marker list can be edited in the Markers enumeration - Content browser: Contents/ARToolkit/Blueprints/Markers
  • Cleaning up at the end.

Level blueprint
===================

  • Transformation of the “chairs” based on marker positions / rotations.

CubeHiro, CubeKanji
===================

  • Simple static meshes - Feel free to change them!

Cheers,
Adam

#16

I have just recognized your new post. Thank you for posting the actual state of your work.
I tried it immediately. As I do not have a printer at home I used a mobile device to display the Hiro marker. Although the mobile device display produced an overexposed image for the camera, your example project was still able to recognize the marker if it was close enough to the camera (and not the whole display of my phone was white on camera’s image).
Tomorrow I will have a look at it with a paper marker.
It is very promising, keep up the good work.

#17

Really great job,and I truly appreciate your time and effort :slight_smile:
Thanks for this wonderful project.

#18

I’ve added NFT (Natural Feature Tracking) to the upcoming plugin. Generally any image can be used as a “marker”.
NFT is also working on Android.

Blueprint used in the video:

bp_ar.jpg
bp_ar.jpg1541×1106 338 KB

#19

The NFT based solution seems even better and Android support is also important for us too. :slight_smile:

#20

This is super awesome. Why no love for mac and iOS ? I’d really love to test this out.