How do I get the picture to stick to the wall？

Question

i'm using arkit and realitykit,first load a room ,then put the picture on the wall ,but how to let it stick on the wall? how to get the right rotation of the picture? There are four possibilities,wall's front: z+, z-, x+, x-,

Answer 1

SceneKit

Answering your first question, to make the node stick to a particular place you need to just keep a desired location and add the node, it should not disappear unless you remove it. It may move up and down a little when you change phone orientation and position, but not much

node.position = SCNVector3(x, y, z)

If you are using touches to define the location, I would recommend you considering sceneView.hitTest() and touchesBegan() functions

As for rotation, you simply can use Euler angles on a node of your interest along any axis, and in any needed direction

node.eulerAngles.x = -.pi / 2

Also, I would highly recommend you “App Development with Swift” by Apple Education, 2019. In this book there is a whole chapter on ARKit, besides answering your questions, it has numerous useful techniques and ideas

Here you can find the implementation of the end-of-the-chapter guided project from the App Development with Swift book, but doing it yourself would be much more useful

RealityKit

As given here, you can use one of two options for rotation

let boxAnchor = try! Experience.loadBox()

boxAnchor.steelBox?.orientation = simd_quatf(angle: .pi/4, axis: [0, 0, 1]) 

For angle property you insert how much you want to rotate the object in radians, and for axis property you select the axis that you want to rotate around.

Another option is to use transform

boxAnchor.steelBox?.transform = Transform(pitch: 0, yaw: 0, roll: .pi/4) 

Roll, pitch, and yaw represent rotation along a particular axis, more is written here.

To change the position you can again use transform's translation

steelBox.transform.translation = [0, 0, -0.5]

This will translate your object according to the given parameters. This function heavily relies on affine transforms.

Regarding transformations w.r.t. other objects

Nodes in SceneKit as well as Entities in RealityKit are transformed and rotated with respect to the parent Node or Entity.

So, in your case, you have a big model where you plan to put smaller objects.

You have two options, either use touches detection on the big model (house) and them manually calculate where the object should be placed, which may be pretty cumbersome.

Another option is to add small transparent planes at predefined positions and then adding touches detection on them. This way you can omit calculation of where the smaller object should be placed

Good luck!