I am making an android application with the use of ARCore and I need to know if the camera is currently facing the object and where the object is in relation to the camera, e.g. is it behind the camera, on the left, on the right, so that you can direct the user how to turn to be in front of the promise. I have access to rotatnio, translation and there is a compose method but I don't know what it's doing.
method compose
Vector3 cameraPosition = camera.getWorldPosition();
Vector3 objPosition = object.getWorldPosition();
Vector3 direction = Vector3.subtract(objPosition, objPcameraPositionosition);
Quaternion lookRotation = Quaternion.lookRotation(direction, Vector3.up());
The above gives you the quaternion which needs to be applied to the camera to make it face the object. You can subtract this from the camera's current world rotation to get the difference and notify the user accordingly.
Related
When I add a new node with ARKit (ARSKView), the object is positioned base on the device camera. So if your phone is facing down or tilted, the object will be in that direction as well. How can I instead place the object base on the horizon?
For that, right after a new node's creation, use a worldOrientation instance property that controls the node's orientation relative to the scene's world coordinate space.
var worldOrientation: SCNQuaternion { get set }
This quaternion isolates the rotational aspect of the node's worldTransform matrix, which in turn is the conversion of the node's transform from local space to the scene's world coordinate space. That is, it expresses the difference in axis and angle of rotation between the node and the scene's rootNode.
let worldOrientation = sceneView.scene.rootNode.worldOrientation
yourNode.rotation = worldOrientation /* X, Y, Z, W components */
P.S. (as you updated your question) :
If you're using SpriteKit, 2D sprites you spawn in ARSKView are always face the camera. So, if the camera moves around a definite point of real scene, all the sprites must be rotated about their pivot point, still facing a camera.
Nothing can prevent you from using SceneKit and SpriteKit together.
I am trying to find distance between iOS device's front-facing camera and user's face in the real world.
So far, I have tried ARKit/SceneKit, and using ARFaceAnchor I am able to detect user's face distance from camera; but it works only in close proximity (up to about 88 cm). My application requires face distance detection up to 200 cms.
I am assuming this could be achieved without the use of trueDepth data (which is being used in ARFaceAnchor).
Can you put me in the right direction?
In order to get the distance between the device and the user's face you should convert position of the detected user's face into camera's coordinate system. To do this, you will have to use the convertPosition method from SceneKit to switch coordinate space, from face coordinate space to camera coordinate space.
let positionInCameraSpace = theFaceNode.convertPosition(pointInFaceCoordinateSpace, to: yourARSceneView.pointOfView)
theFaceNode is the SCNNode created by ARKit representing the user's face. The pointOfView property of your ARSCNView returns the node from which the scene is viewed, basically the camera.
pointInFaceCoordinateSpace could be any vertices of the face mesh or just the position of theFaceNode (which is the origin of the face coordinate system). Here, positionInCameraSpace is a SCNVector3, representing the position of the point you gave, in camera coordinate space. Then you can get the distance between the point and the camera using the x,y and z value of this SCNVector3 (expressed in meters).
these are some links that may help you :
-Distance between face and camera using ARKit
-https://github.com/evermeer/EVFaceTracker
-https://developer.apple.com/documentation/arkit/arfacetrackingconfiguration
-How to measure device distance from face with help of ARKit in iOS?
I'm running face landmark detection using the front-facing camera on iPhone X, and am trying very hard to get 3D points of face landmarks (VNFaceLandmarkRegion2D gives image coordinates X, Y only).
I've been trying to use either the ARSCNView.hitTest or ARFrame.hitTest, but am so far unsuccessful. I think my error may be in converting the initial landmark points to the correct coordinate system. I've tried quite a few permutations, but currently based on my research this is what I've come up with:
let point = CGPoint(x: landmarkPt.x * faceBounds.width + faceBounds.origin.x, y: (1.0 - landmarkPt.y) * faceBounds.height + faceBounds.origin.y)
let screenPoint = CGPoint(point.x * view.bounds.width, point.y * view.bounds.height)
frame.hitTest(screenPoint, types: ARHitTestResult.ResultType.featurePoint)
I've also tried to do
let newPoint = CGPoint(x: point.x, y: 1.0 - point.y)
after the conversion, but nothing seems to work. My frame.hitTest result is always empty. Am I missing anything in the conversion?
Does the front-facing camera add another level to this? (I also tried inverting the initial X value at one point, in case the coordinate system was being mirrored). It also seems to me that the initial landmark normalizedPoints are sometimes negative and also sometimes greater than 1.0, which doesn't make any sense to me. I'm using ARSession.currentFrame?.capturedImage to capture the frame of the front-facing camera, if that's important.
Any help would be very, very appreciated, thanks so much!
-- SOLVED --
For anyone with similar issues:
I am finally getting hit test results!
for point in observation.landmarks?.allPoints?.pointsInImage(imageSize: sceneView.bounds.size) {
let result = ARSCNView.hitTest(point, options: [ARSCNHitTestOption.rootNode: faceNode)
}
I use the face geometry as an occlusion node attached to the face node.
Thanks Rickster!
You're using ARFaceTrackingConfiguration, correct? In that case, the featurePoint hit test type won't help you, because feature points are part of world tracking, not face tracking... in fact, just about all the ARKit hit testing machinery is specific to world tracking, and not useful to face tracking.
What you can do instead is make use of the face mesh (ARFaceGeometry) and face pose tracking (ARFaceAnchor) to work your way from a 2D image point to a 3D world-space (or camera-space) point. There's at least a couple paths you could go down for that:
If you're already using SceneKit, you can use SceneKit's hit testing instead of ARKit's. (That is, you're hit testing against "virtual" geometry modeled in SceneKit, not against a sparse estimate of the real-world environment modeled by ARKit. In this case, the "virtual" geometry of the face mesh comes into SceneKit via ARKit.) That is, you want ARSCNView.hitTest(_:options:) (inherited from SCNSceneRenderer), not hitTest(_:types:). Of course, this means you'll need to be using ARSCNFaceGeometry to visualize the face mesh in your scene, and ARSCNView's node/anchor mapping to make it track the face pose (though if you want the video image to show through, you can make the mesh transparent) — otherwise the SceneKit hit test won't have any SceneKit geometry to find.
If you're not using SceneKit, or for some reason can't put the face mesh into your scene, you have all the information you need to reconstruct a hit test against the face mesh. ARCamera has view and projection matrices that tell you the relationship of your 2D screen projection to 3D world space, ARFaceAnchor tells you where the face is in world space, and ARFaceGeometry tells you where each point is on the face — so it's just a bunch of math to get from a screen point to a face-mesh point and vice versa.
OpenCV docs for solvePnp
In an augmented reality app, I detect the image in the scene so I know imagePoints, but the object I'm looking for (objectPoints) is a virtual marker just stored in memory to search for in the scene, so I don't know where it is in space. The book I'm reading(Mastering OpenCV with Practical Computer Vision Projects ) passes it as if the marker is a 1x1 matrix and it works fine, how? Doesn't solvePnP needs to know the size of the object and its projection so we know who much scale is applied ?
Assuming you're looking for a physical object, you should pass the 3D coordinates of the points on the model which are mapped (by projection) to the 2D points in the image. You can use any reference frame, and the results of the solvePnp will give you the position and orientation of the camera in that reference frame.
If you want to get the object position/orientation in camera space, you can then transform both by the inverse of the transform you got from solvePnp, so that the camera is moved to the origin.
For example, for a cube object of size 2x2x2, the visible corners may be something like: {-1,-1,-1},{1,-1,-1},{1,1,-1}.....
You have to pass the 3D coordinates of the real-world object that you want to map with the image. The scaling and rotation values will depend on the coordinate system that you use.
This is not as difficult as it sounds. See this blog post on head pose estimation. for more details with code.
I have created a Camera class that allows me to move around a scene in first person. The camera has worked just fine until I decided to use it as a location to add something to the 3D world. What I am trying to do is add a cube to the world when I press a mouse button. I want to cube to eventually travel away from the camera, but for now I just want to create it right in front of it. Sometimes it works and sometimes it creates it to one side or the other. It all depends on how much I've rotated and translated the camera.
I am tryinng to find the vector in front of my camera by using the View Matrix like so:
Vector3 inFront = Camera.ViewMatrix.Forward;
I plan to use the vector to add some physics behind the cube and have it travel away from the camera. For now I am just wanting to get a correct Vector.
I know you normally draw thing in the world using the WorldMatrix, but I can't figure out how to convert my ViewMatrix into a WorldMatrix. Still learing :-)
What am I doing wrong?
-Scott
First of all, there is no real difference between a "World Matrix" and a "View Matrix", they are both transformation matrices and the distinction is somewhat arbitrary. Some systems even combine the two (OpenGL simply has a "ModelView" matrix).
Traditionally the "world matrix" is used to move individual models from "model space" to "world space". Then the "view matrix" is used to move all the models from world space into their relative positions in front of the camera (which, in effect, "moves the camera"). And finally the "Projection Matrix" converts the 3D positions into their 2D positions on the screen (generally with a perspective projection). Because they are matrices, they can be multiplied together into a single matrix that can transform points in a single step.
First of all, take a look at the properties of the Matrix struct.
What you also need to realise is that Matrix.Forward returns a Vector3. A Vector3 can represent either a position or a scalar and a direction. You need two of them to represent a position and a direction.
Now, my 3D matrix maths is a bit rusty, but I'm pretty sure that what you want is the Matrix.Translation as the position of the camera in world space. And Matrix.Forward as the forward direction of the camera in world space.
Unless your camera/view matrix is performing a scaling operation on the world (and really it shouldn't), then the Vector3 you get back from Matrix.Forward will have unit length - in other words just a direction (no scalar). Use this to give a direction to move your object in.
I assume you have to location of the camera. Have you tried something like this (I haven't done Matrix/Vector math in a few years so this might be off):
float scalar = 10; // how far away from the camera you want to move the object
Vector3 camPos = ???; // supplied from somewhere elese
Vector3 inFront = Camera.ViewMatrix.Forward;
Vector3 newPos = camPos + inFront * scalar;