Click here to Skip to main content
14,423,188 members

Augmented Reality with the ArcGIS Runtime SDK for iOS

16 Dec 2019CPOL
This article will show how to use the Esri's ArcGIS Runtime SDK and Toolkit for iOS to build a compelling augmented reality experience.

This article is in the Product Showcase section for our sponsors at CodeProject. These articles are intended to provide you with information on products and services that we consider useful and of value to developers.


Augmented Reality (AR) experiences are designed to "augment" the physical world with virtual content. That means showing virtual content on top of a device's camera feed. As the device is moved around, that virtual content respects the real-world scale, position, and orientation of the camera's view. The ArcGIS Runtime SDK for iOS and the ArcGIS Runtime Toolkit for iOS from Esri together provide a simplified approach to developing AR solutions that overlay maps and geographic data on top of a live camera feed. Users can feel like they are viewing digital mapping content in the real world.

In this article, we'll learn how to give users that AR map experience. But first, some terminology: in Runtime parlance, a Scene is a description of a 3D "Map" containing potentially many types of 3D geographic data. A Runtime SceneView is a UI component used to display that Scene to the user. When used in conjunction with the ArcGIS Toolkit, a SceneView can quickly and easily be turned into an AR experience to display 3D geographic data as virtual content on top of a camera feed.

Before getting started, be sure to check out this link to get signed up for a free ArcGIS for Developers subscription.

You will need XCode 10.2 or later and a device running iOS 11.0 or later.

There are links at the end if you want more details on anything presented here.


In order to use the information and code presented in this article you will need to install a couple of items:

  • ArcGIS Runtime SDK for iOS - A modern, high-performance, mapping API that can be used in both Swift and Objective-C.
  • ArcGIS Runtime Toolkit for iOS - Open-source code containing a collection of components that will simplify your iOS app development with ArcGIS Runtime.  This includes the AR component we will be showcasing here.

The AR Toolkit component allows you to build applications using three common AR patterns:

        World-scale – A kind of AR scenario where scene content is rendered exactly where it would be in the physical world. This is used in scenarios ranging from viewing hidden infrastructure to displaying waypoints for navigation. In AR, the real world, rather than a basemap, provides the context for your GIS data.

        Tabletop – A kind of AR scenario where scene content is anchored to a physical surface, as if it were a 3D-printed model. You can walk around the tabletop and view the scene from different angles.  The origin camera in a tabletop scenario is usually at the lowest point on the scene.

        Flyover – Flyover AR is a kind of AR scenario that allows you to explore a scene using your device as a window into the virtual world. A typical flyover AR scenario will start with the scene’s virtual camera positioned over an area of interest. You can walk "through" the data and reorient the device to focus on specific content in the scene.  The origin camera in a flyover scenario is usually above the tallest content in the scene.

The AR toolkit component is comprised of one class: ArcGISARView. This is a subclass of UIView that contains the functionality needed to display an AR experience in your application. It uses ARKit, Apple's augmented reality framework to display the live camera feed and handle real-world tracking and synchronization with the Runtime SDK's AGSSceneView. The ArcGISARView is responsible for starting and managing an ARKit session. It uses an AGSLocationDataSource (a class encapsulating device location information and updates) for getting an initial GPS location and when continuous GPS tracking is required.

Features of ArcGISARView:

  • Allows display of the live camera feed
  • Manages ARKit ARSession lifecycle
  • Tracks user location and device orientation through a combination of ARKit and the device GPS
  • Provides access to an AGSSceneView to display your GIS 3D data over the live camera feed
  • ARScreenToLocation method to convert a screen point to a real-world coordinate
  • Easy access to all ARKit and AGSLocationDataSource delegate methods

The steps needed to implement each of the three AR patterns are similar and will be detailed below.

Information on installing the ArcGIS Runtime SDK can be found here.

Information on incorporting the Toolkit into your project can be found here.

The rest of the article assumes you have installed the SDK, cloned or forked the Toolkit repo and set up your project to incorporate both.  At the bottom of the article there is a link to a GitHub repo containing the code presented here.

Using the code

Creating an AR-enabled application is quite simple using the aformentioned SDK and Toolkit.  The basic steps are:

  • Add the ArcGISARView as a sub-view of your application's view controller view.  This can be accomplished either in code or via a Storyboard/.xib file.
  • Add the required entries to your application's plist file (for the camera and GPS).
  • Create and set an AGSScene on the ArcGISARView.sceneView.scene property (the AGSScene references the 3D data you want to display over the live camera feed).
  • Set a location data source on the ArcGISARView (if you want to track the device location).  More on this later...
  • Call ArcGISARView.startTracking(_) to begin tracking location and device motion when your application is ready to begin it's AR session.

All three AR patterns use the same basic steps.  The differences are in how certain properties of the ArcGISARView are set.  These properties are:

  • originCamera - the initial camera position for the view; used to specify a location when the data is not at your current real-world location.
  • translationFactor - specifies how many meters the camera will move for each meter moved in the real world.  Used in Tabletop and Flyover scenarios.
  • locationDataSource - this specifies the data source used to get location updates.  The ArcGIS Runtime SDK includes AGSCLLocationDataSource which is a location data source that uses CoreLocation to generate location updates.  If you're not interested in location updates (for example in flyover and table top scenarios), this property can be nil.
  • startTracking(_ locationTrackingMode: ARLocationTrackingMode)- the locationTrackingMode argument denotes how you want to use the location data source.

    There are three options:

    • .ignore: this ignores the location data source updates completely
    • .initial: uses only the first location data source location update
    • .continuous: use all location updates


We'll start by creating a basic World-scale AR experience.  This will display a web scene containing the Imagery base map and a feature layer that you can view by moving your device to see the data around you.  The web scene also includes an elevation source to display the data at it's real-world altitude.

The first thing we need to do is add an ArcGISARView to your application.  This can be done in storyboards/.xib files or in code.  First, define the ArcGISARView variable and add it to your view controller's view:

// Creates an ArcGISARView and specifies we want to view the live camera feed.
let arView = ArcGISARView(renderVideoFeed: true)
override func viewDidLoad() {

    // Add the ArcGISARView to our view and set up constraints.
    arView.translatesAutoresizingMaskIntoConstraints = false
        arView.leadingAnchor.constraint(equalTo: view.leadingAnchor),
        arView.trailingAnchor.constraint(equalTo: view.trailingAnchor),
        arView.topAnchor.constraint(equalTo: view.topAnchor),
        arView.bottomAnchor.constraint(equalTo: view.bottomAnchor)

Next you'll need to add the required privacy keys for the camera and device location to your application's plist file:

<string>For displaying the live camera feed</string>

<string>For determining your current location</string>


In order to load and display data, create an AGSScene.  We'll use a web scene containing a base map and a feature layer.  As the feature data is location-specific, you probably won't see any of it at your location and would need to add your own data.  Add this code to your viewDidLoad method:

// Create our scene and set it on the arView.sceneView.
let scene = AGSScene(url: URL(string: "")!)!
arView.sceneView.scene = scene

Then create a location data source and set it on arView.  We're using AGSCLLocationDataSource included in the ArcGISRuntime SDK, to provide location information using Apple's CoreLocation framework. After setting the scene, add this line:

arView.locationDataSource = AGSCLLocationDataSource()

The last step is to call startTracking on arView to initiate the location and device motion tracking using ARKit.  This is accomplished in your view controller's viewDidAppear method, since we don't need to start tracking location and motion until the view is dislayed.  You should also stop tracking when the view is no longer visible (in viewDidDisappear).

override func viewDidAppear(_ animated: Bool) {
override func viewDidDisappear(_ animated: Bool) {

We used .continous above because we want to continually track the device location using the GPS.  This is common for World-scale AR experiences.  For Tabletop and Flyover experiences, you would use either .initial, to get only the initial device location or .ignore, when the device location is not needed (for example when setting an origin camera).

At this point you can build and run the app.  The display of the base map will look similar to this, but at your geographic location.  The tree data is location-specific, but by adding your own data you can create a rich experience with very little code.  Rotating, tilting, and panning the device will cause the display of the scene to change accordingly.

World-scale screen shot


Now that we've created a basic AR experience for World-scale scenarios, let's move on to a Tabletop scenario.  In Tabletop AR, data is anchored to a physical surface, as if it were a 3D-printed model. You can walk around the tabletop and view the scene from different angles.

A couple of the differences with Tabletop as compared to World-scale are the ability to explore more than just your immediate surroundings and not using the device's real-world location to determine where in the virtual world to position the camera.

ArcGISARView's translationFactor property allows for more movement in the virtual world than that represented in the real world.  The default value for translationFactor is 1.0, which means moving your device 1 meter in the real word moves the camera 1 meter in the virtual world.  Setting the translationFactor to 500.0 means that for every meter in the real world the device is moved the camera will move 500.0 meters in the virtual world. We'll also set an origin camera, which represents the initial location and orientation of the camera, instead of using the device's real-world location.

We'll use some different data this time, so we'll create a new scene with no base map, a layer representing data of Yosemite National Park, and an elevation source (using an AGSScene extension).

// Create the scene.
let scene = AGSScene()

// Create the Yosemite layer.
let layer = AGSIntegratedMeshLayer(url: URL(string: "")!)

// Set the scene on the arView.sceneView.
arView.sceneView.scene = scene

// Create and set the origin camera.
let camera = AGSCamera(latitude: 37.730776, longitude: -119.611843, altitude: 1213.852173, heading: 0, pitch: 90.0, roll: 0)
arView.originCamera = camera

// Set translationFactor.
arView.translationFactor = 18000

Here's the AGSScene extension which will add an elevation source to your scene:

// MARK: AGSScene extension.
extension AGSScene {
    /// Adds an elevation source to the given scene.
    func addElevationSource() {
        let elevationSource = AGSArcGISTiledElevationSource(url: URL(string: "")!)
        let surface = AGSSurface()
        surface.elevationSources = [elevationSource] = "baseSurface"
        surface.isEnabled = true
        surface.backgroundGrid.isVisible = false
        surface.navigationConstraint = .none
        baseSurface = surface

In viewDidAppear, you can pass .ignore to startTracking to ignore location data:

// For Tabletop AR, ignore location updates.
// Start tracking, but ignore the GPS as we've set an origin camera.

When running the app, if you move the camera around slowly, pointed at the table top or other flat surface you want to anchor the data to, the arView (using ARKit) will automatically detect horizontal planes.  These planes can be used to determine the surface you want to anchor to.  The planes can be visualized using the following lines of code and the Plane class defined below.

In viewDidLoad:

// Set ourself as delegate so we can get ARSCNViewDelegate method calls.
arView.arSCNViewDelegate = self

Next, implement the ARSCNViewDelegate as an extension.  You'll need to import ARKit in order to get the definition for ARSCNViewDelegate.

import ARKit

// MARK: ARSCNViewDelegate
extension ViewController: ARSCNViewDelegate {
    func renderer(_ renderer: SCNSceneRenderer, didAdd node: SCNNode, for anchor: ARAnchor) {
        // Place content only for anchors found by plane detection.
        guard let planeAnchor = anchor as? ARPlaneAnchor else { return }
        // Create a custom object to visualize the plane geometry and extent.
        let plane = Plane(anchor: planeAnchor, in: arView.arSCNView)
        // Add the visualization to the ARKit-managed node so that it tracks
        // changes in the plane anchor as plane estimation continues.
    func renderer(_ renderer: SCNSceneRenderer, didUpdate node: SCNNode, for anchor: ARAnchor) {
        // Update only anchors and nodes set up by renderer(_:didAdd:for:).
        guard let planeAnchor = anchor as? ARPlaneAnchor,
            let plane = node.childNodes.first as? Plane
            else { return }
        // Update extent visualization to the anchor's new bounding rectangle.
        if let extentGeometry = plane.node.geometry as? SCNPlane {
            extentGeometry.width = CGFloat(planeAnchor.extent.x)
            extentGeometry.height = CGFloat(planeAnchor.extent.z)
            plane.node.simdPosition =

Once a plane has been displayed, the user can tap on it and anchor the data to it.  We can make the sceneView semi-transparent to better see the detected planes (in the viewDidLoad method):

// Dim the SceneView until the user taps on a surface.
arView.sceneView.alpha = 0.5

In order to capture the user tap, you want to implement the AGSGeoViewTouchDelegate of the sceneView.  We do this by settting the touchDelegate on the sceneView in viewDidLoad:

// Set ourself as touch delegate so we can get touch events.
arView.sceneView.touchDelegate = self

Then we implement the geoView:didTapAtScreenPoint:mapPoint: method in an extension:

// MARK: AGSGeoViewTouchDelegate
extension ViewController: AGSGeoViewTouchDelegate {
    public func geoView(_ geoView: AGSGeoView, didTapAtScreenPoint screenPoint: CGPoint, mapPoint: AGSPoint) {
        // Place the scene at the given point by setting the initial transformation.
        if arView.setInitialTransformation(using: screenPoint) {
            // Show the SceneView now that the user has tapped on the surface.
            UIView.animate(withDuration: 0.5) { [weak self] in
                self?.arView.sceneView.alpha = 1.0

The arView.setInitialTransformation method will take the tapped screen point and determine if an ARKit plane intersects the screen point; if one does, it will set the arView.initialTransformation property, which has the effect of anchoring the data to the tapped-on plane.  The above code will also make the sceneView fully opaque.

The Plane class for visualizing ARKit planes:

/// Helper class to visualize a plane found by ARKit
class Plane: SCNNode {
    let node: SCNNode
    init(anchor: ARPlaneAnchor, in sceneView: ARSCNView) {
        // Create a node to visualize the plane's bounding rectangle.
        let extent = SCNPlane(width: CGFloat(anchor.extent.x), height: CGFloat(anchor.extent.z))
        node = SCNNode(geometry: extent)
        node.simdPosition =
        // `SCNPlane` is vertically oriented in its local coordinate space, so
        // rotate it to match the orientation of `ARPlaneAnchor`.
        node.eulerAngles.x = -.pi / 2


        node.opacity = 0.6
        guard let material = node.geometry?.firstMaterial
            else { fatalError("SCNPlane always has one material") }
        material.diffuse.contents = UIColor.white

        // Add the plane node as child node so they appear in the scene.
    required init?(coder aDecoder: NSCoder) {
        fatalError("init(coder:) has not been implemented")

Run the app and slowly move the device around your table top to find a plane; once one is found, tap on it to place the data.  You can then move the device around the table to view the data from all angles.

Tabletop screenshot


The last AR scenario we will cover is Flyover.  Flyover AR is a kind of AR scenario that allows you to explore a scene using your device as a window into the virtual world.  We'll start out the same way as Tabletop, by creating a new scene with no base map, a layer representing data along the US-Mexican border, and an elevation source:

// Create the scene.
let scene = AGSScene()

// Create the border layer.
let layer = AGSIntegratedMeshLayer(url: URL(string: "")!)

// Add layer to the scene's operational layers array.

// Set the scene on the arView.sceneView.
arView.sceneView.scene = scene

// Create and set the origin camera.
let camera = AGSCamera(latitude: 32.533664, longitude: -116.924699, altitude: 126.349876, heading: 0, pitch: 90.0, roll: 0)
arView.originCamera = camera

// Set translationFactor.
arView.translationFactor = 1000

We keep the arView.startTracking(.ignore) method the same as for the Tabletop scenario.

// For Tabletop and Flyover AR, ignore location updates.
// Start tracking, but ignore the GPS as we've set an origin camera.

Building and running the app at this point allows you to view the data from a higher altitude and "fly" over the data to view the data in a different way.

Flyover screen shot

Points of Interest

One of the tricky parts of building a full-featured AR experience on a mobile device is the relative inacurracy of the device GPS location and heading.  The accuracy will vary in importance depending on the type of AR scenario.  For Tabletop and Flyover, the accuracy is usually not important, as in most of those cases you will be setting the initial location directly.  For World-scale AR, this can be very important, especially if you are trying to locate real-world features based on their digital counterparts.  Some type of calibration, where the location and heading are adjusted to match the real-world location and heading can be necessary.  The ArcGIS Runtime Toolkit for iOS has an example of a calibration view and workflow to handle those inaccuracies than you can incorporate into your application.  See the link below on the Toolkit's AR example for more information.

More information:

  • Apple ARKit
  • ArcGIS Toolkit for iOS contains an AR example which allows users to choose from several different AR scenarios.  More information on the Toolkit's AR component and example can be found here.
  • For more information and detail on using the ArcGISRuntime SDK and the ArcGIS Runtime Toolkit to build compelling, full-featured applications, see the ArcGIS Runtime SDK Guide topic: Display scenes in augmented reality.
  • A GitHub repo containing a fully-functional app demonstrating the code in this article can be found here.


This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)


About the Author

Mark Dostal
Software Developer Esri
United States United States
Mark is an iOS and Mac developer at Esri, the global market leader in GIS.
He’s worked in Software Development since 1988, published his first Mac app in 1989, attended his first WWDC in the early ’90s, worked for local educational software company MECC on The Oregon Trail, contributed to Esri’s first Mac app, which won “Best in Show” at the ’95 MacWorld conference, worked on Esri’s first iOS app, which had over 900k downloads, and is currently helping developers from all over the world create compelling and immersive mapping applications.

Comments and Discussions

-- There are no messages in this forum --
Posted 16 Dec 2019


2 bookmarked