RealityKit – Convert CGPoint to a SIMD<Float> - swift

I am using RealityKit and when I tap on the screen, I just want to add a box on that place where I tapped. It can be on the middle of the screen or anywhere. I will adjust Z axis so the object is placed 1 meter away from me. But I am having a hard time converting CGPoint to relevant RealityKit coordinates.
#objc func handleTap(_ recognizer: UITapGestureRecognizer) {
guard let view = view else { return }
let location = recognizer.location(in: view)
// location to SIMD3<Float> so I can create an anchor
}
Any ideas?

Model on a detected plane
CGPoint is a XY-point on iPhone's screen, so there's no need to convert it to XYZ-point. If you need a 3D point on a detected plane for accomodating your model, all you have to do is to generate an object from ARRaycastResult (or ARHitTestResult in case you're using ARKit+SceneKit).
import RealityKit
import ARKit
var arView = ARView(frame: .zero)
#objc func tappingScreen(_ sender: UITapGestureRecognizer) {
let results: [ARRaycastResult] = arView.raycast(from: arView.center,
allowing: .estimatedPlane,
alignment: .horizontal)
if let result: ARRaycastResult = results.first {
let model = ModelEntity(mesh: .generateSphere(radius: 0.02))
let anchor = AnchorEntity(world: result.worldTransform)
anchor.addChild(model)
arView.scene.anchors.append(anchor)
}
}
Model at camera's position
Creating a model exactly where ARCamera is located in the current frame is super easy:
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
let model = ModelEntity(mesh: .generateSphere(radius: 0.02))
let anchor = AnchorEntity(world: arView.cameraTransform.matrix)
anchor.addChild(model)
model.position.z = -0.5 // 50 cm offset
arView.scene.anchors.append(anchor)
}
Also, you can multiply camera's transform by desired offset:
var translation = matrix_identity_float4x4
translation.columns.3.z = -0.5 // 50 cm offset
// Multiplying camera matrix by offset matrix
let transform = simd_mul(arView.cameraTransform.matrix, translation)
model.transform.matrix = transform
Additional info
This post is also helpful.

Related

Changing the angle of a SKSpriteNode during a rotation

I am trying to make a drifting game. In order for the car to drift, the car (when turning) needs to be at an angle. I have tried rotation, however this conflicts with the code I already have for turning the car. Here is my code, any help?
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
if let touch = touches.first {
let location = touch.previousLocation(in: self)
let position = touch.location(in: self)
let node = self.nodes(at: location).first
if position.x < 0 {
let rotate = SKAction.repeatForever(SKAction.rotate(byAngle: CGFloat(M_PI), duration: 0.8))
car.run(rotate, withKey: "rotating")
} else {
let rotate = SKAction.repeatForever(SKAction.rotate(byAngle: CGFloat(-M_PI), duration: 0.8))
car.run(rotate, withKey: "rotating")
}
}
}
override func update(_ currentTime: TimeInterval) {
// Called before each frame is rendered
car.position = CGPoint(x:car.position.x + cos(car.zRotation) * 3.0,y:car.position.y + sin(car.zRotation) * 3.0)
}
}
This code currently turns the car without adding any angle, or "drifting" effect.
one approach is you could nest your car inside another SKNode as a container. apply your steering rotation to the car node, as you're doing it now. then apply the drift rotation to the container node. the result effect will be the sum of the two.
//embed car inside a SKNode container so you can apply different rotations to each
let car = SKShapeNode(ellipseOf: CGSize(width: 20, height: 40)) //change to how you draw your car
let drift_container = SKNode()
drift_container.addChild(car) //embed car inside the container
self.addChild(drift_container) //`self` here is a SKScene
//apply angle rotation to the container
func drift(byAngle angle:CGFloat) {
let rotate = SKAction.rotate(toAngle: angle, duration: 0.2)
drift_container.run(rotate)
}
then in update be sure to update drift_container.position instead of the car's position

Implement a crosshair kind behaviour in RealityKit

What I want to achieve: Attach a sphere to the camera position (so that it always stay at the center of the screen as the device move) and detect when it is on top of other AR objects - to trigger other actions/behaviour on the AR objects.
Approach: I have created the sphere and attached to the center of the screen as shown below
#IBOutlet var arView: ARView!
override func viewDidLoad() {
super.viewDidLoad()
let mesh = MeshResource.generateSphere(radius: 0.1)
let sphere = ModelEntity(mesh: mesh)
let anchor = AnchorEntity(.camera)
sphere.setParent(anchor)
arView.scene.addAnchor(anchor)
sphere.transform.translation.z = -0.75
}
Next step, perform a hittest or a raycast in session(_:didUpdate:)
let results = arView.hitTest(CGPoint(x: 0.5, y: 0.5), query: .all, mask: .default)
//normalised center ; 2D position of the camera (our sphere) in the view’s coordinate system
But I am constantly getting ground plane as my result with this approach. Is there something I am missing or there is a different approach to achieving this
Note: Just in case there is something wrong I have created my basic scene as I want to track an image and add content on top of the image marker in Reality Composer and using the .rcproject in Xcode also have enabled collision property for all the overlaid items.
Try the following solution:
import ARKit
import RealityKit
class ViewController: UIViewController {
#IBOutlet var arView: ARView!
var sphere: ModelEntity?
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
let touch = arView.center
let results: [CollisionCastHit] = arView.hitTest(touch)
if let result: CollisionCastHit = results.first {
if result.entity.name == "Cube" && sphere?.isAnchored == true {
print("BOOM!")
}
}
}
override func viewDidLoad() {
super.viewDidLoad()
// Crosshair
let mesh01 = MeshResource.generateSphere(radius: 0.01)
sphere = ModelEntity(mesh: mesh01)
sphere?.transform.translation.z = -0.15
let cameraAnchor = AnchorEntity(.camera)
sphere?.setParent(cameraAnchor)
arView.scene.addAnchor(cameraAnchor)
// Model for collision
let mesh02 = MeshResource.generateBox(size: 0.3)
let box = ModelEntity(mesh: mesh02, materials: [SimpleMaterial()])
box.generateCollisionShapes(recursive: true)
box.name = "Cube"
let planeAnchor = AnchorEntity(.plane(.any,
classification: .any,
minimumBounds: [0.2, 0.2]))
box.setParent(planeAnchor)
arView.scene.addAnchor(planeAnchor)
}
}

how to find centre of a plane found in a hit test ARKit SceneKit

In an image detection app, the image is recognised, then an opaque overlay plane is created so when the user taps on the screen a hit test finds the overlay plane, and a new object can be created. But I want to position the object exactly at the centre of the underlying image. How can I get it to be always at the centre of the image / plane, and to have the same orientation. Can this be got from a hit test result? Thanks for any advice!
#objc func handleScreenTap(sender: UITapGestureRecognizer) {
let tappedSceneView = sender.view as! ARSCNView
let tapLocation = sender.location(in: tappedSceneView)
let planeIntersections = tappedSceneView.hitTest(tapLocation, types: [.estimatedHorizontalPlane, .estimatedVerticalPlane])
if !planeIntersections.isEmpty {
addSceneAtPositionOnPlane(hitTestResult: planeIntersections.first!)
}
func addSceneAtPositionOnPlane(hitTestResult: ARHitTestResult) {
let transform = hitTestResult.worldTransform
let positionColumn = transform.columns.3
let initialPosition = SCNVector3(positionColumn.x,
positionColumn.y,
positionColumn.z)
let node = self.createScene(for: initialPosition)
sceneView.scene.rootNode.addChildNode(node)
}
func createScene(for position: SCNVector3) -> SCNNode {
let box = SCNNode(geometry: SCNBox(width: 0.1, //x
height: 0.1, //y
length: 0.1, //z
chamferRadius: 0))
box.geometry?.firstMaterial?.diffuse.contents = UIColor.red
box.geometry?.firstMaterial?.isDoubleSided = true
box.opacity = 0.8
box.position = position
return box
}
if you already added a SCNNode to render a plane on top of the detected image, then you could just use the SceneKit hitTest method that returns a SceneKit node vs. trying to hit test against ARKit geometry.
Once you have the plane you added to the scene you can just add your new geometry as a child of that node.
Here is an example where once the image is detected a plane is drawn on top of it, then when the user clicks on the plane a box is added as a child, the box will then follow the tracked image around and have the correct position and orientation.
import UIKit
import SceneKit
import ARKit
class ViewController: UIViewController, ARSCNViewDelegate {
#IBOutlet var sceneView: ARSCNView!
override func viewDidLoad() {
super.viewDidLoad()
sceneView.delegate = self
let tapGesture = UITapGestureRecognizer(target: self, action: #selector(onTap))
sceneView.addGestureRecognizer(tapGesture)
}
#objc func onTap(_ recognizer: UITapGestureRecognizer) {
let point = recognizer.location(in: sceneView)
guard let hit = sceneView.hitTest(point, options: nil).first else {
return
}
let box = SCNBox(width: 0.02, height: 0.02, length: 0.02, chamferRadius: 0)
let node = SCNNode(geometry: box)
node.position = SCNVector3(0, 0, 0.01)
box.materials.first?.diffuse.contents = UIColor.red
hit.node.addChildNode(node)
}
override func viewWillAppear(_ animated: Bool) {
super.viewWillAppear(animated)
let configuration = ARWorldTrackingConfiguration()
guard let images = ARReferenceImage.referenceImages(inGroupNamed: "ARTest", bundle: nil) else {
return
}
configuration.detectionImages = images
configuration.maximumNumberOfTrackedImages = 1
sceneView.session.run(configuration)
}
func renderer(_ renderer: SCNSceneRenderer, didAdd node: SCNNode, for anchor: ARAnchor) {
guard let imageAnchor = anchor as? ARImageAnchor else {
return
}
let size = imageAnchor.referenceImage.physicalSize
let plane = SCNPlane(width: size.width, height: size.height)
let planeNode = SCNNode(geometry: plane)
planeNode.eulerAngles.x = -Float.pi / 2
planeNode.opacity = 0.9
node.addChildNode(planeNode)
}
override func viewWillDisappear(_ animated: Bool) {
super.viewWillDisappear(animated)
sceneView.session.pause()
}
}

Swift: ARKit place a rectangle based on 2 nodes

My idea is that i want to place 2 sphere nodes at selected locations. From that point i basically want to draw a rectangle that will adjust the height with a slider. Basically that means that the 2 spheres will represent all 4 corners in the beginning. But when testing the code i use 1 meter height as a test.
The problem i have is that i cant seem to place the rectangle at the correct location as illustrated in the image below:
the rectangle in the image has a higher y-point than the points, it's rotated slightly and its center point is above node 2 and not in between the nodes. I don't want to use node.rotation as it wont work dynamically.
This is the code i use to place the 2 nodes and draw + add the rectangle.
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
guard let location = touches.first?.location(in: sceneView) else {return}
let hitTest = sceneView.hitTest(location, types: [ARHitTestResult.ResultType.featurePoint])
guard let result = hitTest.last else {return}
// Converts the matrix_float4x4 to an SCNMatrix4 to be used with SceneKit
let transform = SCNMatrix4.init(result.worldTransform)
// Creates an SCNVector3 with certain indexes in the matrix
let vector = SCNVector3Make(transform.m41, transform.m42, transform.m43)
let node = addSphere(withPosition: vector)
nodeArray.append(node)
sceneView.scene.rootNode.addChildNode(node)
if nodeArray.count == 2 {
let node1 = sceneView.scene.rootNode.childNodes[0]
let node2 = sceneView.scene.rootNode.childNodes[1]
let bezeierPath = UIBezierPath()
bezeierPath.lineWidth = 0.01
bezeierPath.move(to: CGPoint(x: CGFloat(node1.position.x), y: CGFloat(node1.position.y)))
bezeierPath.addLine(to: CGPoint(x: CGFloat(node2.position.x), y: CGFloat(node2.position.y)))
bezeierPath.addLine(to: CGPoint(x: CGFloat(node2.position.x), y: CGFloat(node2.position.y+1.0)))
bezeierPath.addLine(to: CGPoint(x: CGFloat(node1.position.x), y: CGFloat(node1.position.y+1.0)))
bezeierPath.close()
bezeierPath.fill()
let shape = SCNShape(path: bezeierPath, extrusionDepth: 0.02)
shape.firstMaterial?.diffuse.contents = UIColor.red.withAlphaComponent(0.8)
let node = SCNNode.init(geometry: shape)
node.position = SCNVector3(CGFloat(abs(node1.position.x-node2.position.x)/2), CGFloat(abs((node1.position.y)-(node2.position.y))/2), CGFloat(node1.position.z))
sceneView.scene.rootNode.addChildNode(node)
}
}
Also note that this is not my final code. It will be refactored once i get everything working :).
This is currently working. I used the wrong calculation to set it in the middle of the 2 points.
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
// Check if a window is placed already and fetch location in sceneView
guard debugMode, let location = touches.first?.location(in: sceneView) else {return}
// Fetch targets at the current location
let hitTest = sceneView.hitTest(location, types: [ARHitTestResult.ResultType.featurePoint])
guard let result = hitTest.last else {return}
// Create sphere are position. getPosition() is an extension
createSphere(withPosition: result.getPosition())
// When 2 nodes has been placed, create a window and hide the spheres
if nodeArray.count == 2 {
let firstNode = nodeArray[nodeArray.count-1]
let secondNode = nodeArray[nodeArray.count-2]
firstNode.isHidden = true
secondNode.isHidden = true
createWindow(firstNode: firstNode, secondNode: secondNode)
sceneView.debugOptions = []
}
}
func createSphere(withPosition position: SCNVector3) {
// Create an geometry which you will apply materials to and convert to a node
let object = SCNSphere(radius: 0.01)
let material = SCNMaterial()
material.diffuse.contents = UIColor.red
object.materials = [material]
let node = SCNNode(geometry: object)
node.position = position
nodeArray.append(node)
sceneView.scene.rootNode.addChildNode(node)
}
func createWindow(firstNode: SCNNode, secondNode: SCNNode) {
// Create an geometry which you will apply materials to and convert to a node
let shape = SCNBox(width: CGFloat(abs(firstNode.worldPosition.x-secondNode.worldPosition.x)), height: 0.01, length: 0.02, chamferRadius: 0)
let material = SCNMaterial()
material.diffuse.contents = UIColor.red.withAlphaComponent(0.8)
// Each side of a cube can have different materials
// https://stackoverflow.com/questions/27509092/scnbox-different-colour-or-texture-on-each-face
shape.materials = [material]
let node = SCNNode(geometry: shape)
let minX = min(firstNode.worldPosition.x, secondNode.worldPosition.x)
let maxX = max(firstNode.worldPosition.x, secondNode.worldPosition.x)
// The nodes pivot Y-axis should be split in half of the nodes height, this will cause the node to
// only scale in one direction, when scaling it
// https://stackoverflow.com/questions/42568420/scenekit-understanding-the-pivot-property-of-scnnode/42613002#42613002
node.position = SCNVector3(((maxX-minX)/2)+minX, firstNode.worldPosition.y, firstNode.worldPosition.z)
node.pivot = SCNMatrix4MakeTranslation(0, 0.005, 0)
node.scale = SCNVector3Make(1, -50, 1)
node.name = "window"
sceneView.scene.rootNode.addChildNode(node)
}
Also added:
#objc func resetSceneView() {
// Clear all the nodes
sceneView.scene.rootNode.enumerateHierarchy { (node, stop) in
node.childNodes.forEach({
$0.removeFromParentNode()
})
node.removeFromParentNode()
}
// Reset the session
sceneView.session.pause()
sceneView.debugOptions = [ARSCNDebugOptions.showFeaturePoints]
sceneView.session.run(configuration, options: .resetTracking)
let matrix = sceneView.session.currentFrame!.camera.transform
sceneView.session.setWorldOrigin(relativeTransform: matrix)
nodeArray = []
}
To make sure that the world alignment is reset to your camera position

SCNBox make the front side rotated to the camera in ARKit

I have a simple ARKit app. When the user touches the screen
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
guard let touch = touches.first else { return }
let result = sceneView.hitTest(touch.location(in: sceneView), types: [ARHitTestResult.ResultType.featurePoint])
guard let hitResult = result.last else { return }
let hitTrasform = SCNMatrix4(hitResult.worldTransform)
let hitVector = SCNVector3Make(hitTrasform.m41, hitTrasform.m42, hitTrasform.m43)
createBox(position: hitVector)
}
I add a cube
func createBox(position: SCNVector3) {
let box = SCNBox(width: 0.1, height: 0.1, length: 0.1, chamferRadius: 1)
box.firstMaterial?.diffuse.contents = UIColor.black
let sides = [
UIColor.red, // Front
UIColor.black, // Right
UIColor.black, // Back
UIColor.black, // Left
UIColor.black, // Top
UIColor.black // Bottom
]
let materials = sides.map { (side) -> SCNMaterial in
let material = SCNMaterial()
material.diffuse.contents = side
material.locksAmbientWithDiffuse = true
return material
}
box.materials = materials
let boxNode = SCNNode(geometry: box)
boxNode.position = position
sceneView.scene.rootNode.addChildNode(boxNode)
}
When I add it for the first time, I see the front side and a bit of the right side. if I go around the cube on the other side and add another cube, I see the back side of the cube. So, can I do it so that the user sees only the front side whatever how the user goes around.
Thank you.
You can use SCNBillboardConstraint to make a given node face the camera. Just add boxNode.constraints = [SCNBillboardConstraint()] before adding it to the rootNode.