Augment your reality with ARCore !
Download as PPTX, PDF2 likes951 views
Augmented reality (AR) overlays digital information on the real world. ARCore and ARKit are platforms for building AR apps on Android and iOS respectively. ARCore uses motion tracking and environmental understanding to allow virtual objects to interact with the real world. It was developed by Google as the successor to Tango. ARKit uses visual-inertial odometry and plane detection for positioning virtual objects. While ARKit has better tracking, ARCore enables more reliable tracking recovery and environmental mapping. Both aim to make AR development easier by handling complex computer vision tasks.
Augment your reality with ARCore !
- 1. 1 Augment your reality with ARCore Wajdi Ben Rabah
- 3. #MIXIT18
- 4. Augmented Reality ⬡A way of fusing the real and the virtual world by overlaying digital data on to real-world analogue views. ⬡Anyone with a mobile phone or laptop with built-in video capability can augment and extend their reality with easy-to-use applications. 1/39
- 5. Application areas 2/39 Games Marketing Fashion Social media Health Tourism Industry (cars, boats …) Press Ads
- 6. Virtual Reality ⬡Artificial environment, presented as reality alternative. ⬡Experienced senses: ⬡Hearing. ⬡Sight. 3/39
- 8. VR vs AR ⬡AR: ⬡Combine the real world and virtual data in real time. ⬡Be interactive in real world. ⬡Using a 3D environment. ⬡VR: ⬡Alternative to the real world. ⬡Creation of content. 5/39
- 11. ARCore ⬡Release : Dec’17 ⬡ARCore is a platform for building augmented reality apps on Android. ⬡In simple words, ARCore is meant to enable us to interact with digital content in a physical way. 8/39
- 12. ARCore 9/39
- 13. ARCore ⬡ The fruit of the tango project (beginning from 2014) ⬡ Portable (unlike tango) 10/39
- 14. ARCore ⬡ Motion tracking allows the phone to understand and track its position relative to the world. ⬡ Environmental understanding allows the phone to detect the size and location of flat horizontal surfaces like the ground or a coffee table. ⬡ Light estimation allows the phone to estimate the environment's current lighting conditions. 11/39
- 15. ARCore released ⬡ A reference design for hardware to all OEMs. ⬡ Bigger cost for the OEMs. 12/39
- 16. ARKit 13/39
- 17. ARKit ⬡ VIO system with simple 2D plane detection. ⬡ Tracking position in real world in real time: ⬡ Visual tracking (camera): a pixel is mapped to a point in RW (30fp/s). =>based on distance ⬡ IMU : Gyroscope+Accelerometer (reading 1k/s) =>based on acceleration. ⬡ Result in the Kalman filter: which one is providing the best estimation (ground truth) and publish to ARkit SDK. 14/39
- 18. ARKit ⬡ Dead Reckoning : measure user motion/device movement between IMU readings (a guess) : source of error if it continues without a reset. ⬡ Simple plane detection: “sparse” point cloud 15/39
- 19. ARKit ⬡ How ? ⬡ One lens. ⬡ Normally 2 camera: 2 frames, calculate the distance with the imu Dead reckoning. ⬡ 1 camera : capture one frame then move and capture the second. ⬡ ++Error removal 16/39
- 20. ARKit ⬡ Couple VIO algorithms to sensors. ⬡ Calibrating to eliminate errors. ⬡ Full control of the production chain. 17/39
- 21. Optical calibration ⬡ Geometric calibration : FOV for example ⬡ Photometric calibration : color mapping, matching pixel with real world point ... 18/39
- 23. ARCore vs ARKit ⬡ ARKit has some advantages when it comes to tracking, referring to a person’s phone understanding where it is in 3D space, ⬡ ARCore is better equipped for mapping a person’s surrounding as well as more reliable tracking recovery. 20/39
- 25. Tango ⬡ Tango started out mostly focussed on tracking the motion of the phone in 3D space. Many of the original ideas were aimed at indoor mapping. It was only later that AR & VR became the most popular use-cases. ⬡ Hardware reference design (like hololens) : hardware reference design ⬡ Source code based on what FlyBy done. 22/39
- 26. Sparse mapping Depth perception ...Fish eyes Depth camera Tango 23/39
- 27. Hololens Same software stack for VIO + HPU HPU Holographic Processing Units Software Stack VIO Hololens 24/39
- 28. Vuforia ⬡ Hardware independent ⬡ Cross platform ⬡ Generally you need a marker 25/39
- 29. Wikitude ⬡ Supported Platforms: Android, iOS, Smart Glasses. 26/39
- 30. OpenCV and other open source projects ⬡ Visual tracker only ⬡ some use of sparse maps 27/39
- 31. WEBXR WebXR makes it easier for developers to target all devices, just like when the web browser debuted decades ago for the first time, making it easy to target millions of devices with a single codebase. 28/39
- 32. WEBXR 29/39
- 34. Android if (session == null) { Exception exception = null; String message = null; try { switch (ArCoreApk.getInstance().requestInstall(this, !installRequested)) { case INSTALL_REQUESTED: installRequested = true; return; case INSTALLED: break; } // ARCore requires camera permissions to operate. If we did not yet obtain runtime // permission on Android M and above, now is a good time to ask the user for it. if (!CameraPermissionHelper.hasCameraPermission(this)) { CameraPermissionHelper.requestCameraPermission(this); return; } 30/39
- 35. Android // Create the session. session = new Session(/* context= */ this); } catch (UnavailableArcoreNotInstalledException | UnavailableUserDeclinedInstallationException e) { message = "Please install ARCore"; exception = e; } catch (UnavailableApkTooOldException e) { message = "Please update ARCore"; exception = e; } catch (UnavailableSdkTooOldException e) { message = "Please update this app"; exception = e; } catch (UnavailableDeviceNotCompatibleException e) { message = "This device does not support AR"; exception = e; } catch (Exception e) { message = "Failed to create AR session"; exception = e; } 31/39
- 36. Android @Override public void onSurfaceCreated(GL10 gl, EGLConfig config) { GLES20.glClearColor(0.1f, 0.1f, 0.1f, 1.0f); // Prepare the rendering objects. This involves reading shaders, so may throw an IOException. try { // Create the texture and pass it to ARCore session to be filled during update(). backgroundRenderer.createOnGlThread(/*context=*/ this); planeRenderer.createOnGlThread(/*context=*/ this, "models/trigrid.png"); pointCloudRenderer.createOnGlThread(/*context=*/ this); virtualObject.createOnGlThread(/*context=*/ this, "models/andy.obj", "models/andy.png"); virtualObject.setMaterialProperties(0.0f, 2.0f, 0.5f, 6.0f); virtualObjectShadow.createOnGlThread( /*context=*/ this, "models/andy_shadow.obj", "models/andy_shadow.png"); virtualObjectShadow.setBlendMode(BlendMode.Shadow); virtualObjectShadow.setMaterialProperties(1.0f, 0.0f, 0.0f, 1.0f); } catch (IOException e) { Log.e(TAG, "Failed to read an asset file", e); } } 32/39
- 37. Unity3D ⬡Unity is a platform for creating experiences in 2D and 3D. ⬡SDK : a software development kit for ARCore. 33/39
- 38. ARCore Unity ⬡ Point cloud : Contains a set of observed 3D points and confidence values. ⬡ Arcore device : contains the session and tracking type ⬡ Change to android : build settings ⬡ minimum api 7.0 ⬡ multithread rendering : disabled (cpu) ⬡ xr settings : arcore support ⬡ Download arcore services and proceed to install (via adb for example) 34/39
- 39. Blocks 35/39
- 40. Demo 36/39
- 42. Future of AR ⬡ Depth camera : power consumption, sunlight ⬡ Deep learning : outdoor relocalization ⬡ Large area mapping ⬡ Inputs ⬡ 3D reconstruction : shape of objects in real world : collision/occlusion 38/39
- 43. Future of AR ⬡Limitations: low light/excessive motion/flat surfaces with no textures/lose sensors (camera/gyro/accelerometer)/ plane detection takes some time. ⬡Unite Austin 2017: Cross platform AR apps ⬡Multiplayer ar: tanks network demo 39/39