- Home /

# Can you help me understand this Camera calibration between Kinect and PS Move?

I recently found a piece of code that magically combines PS Move and Kinect coordinates and fixes a displacement issue that I had. It's nice that my problem is solved, but I have no clue how this calibration happened! I had thought that Calibration needs to happen using one of these algorithms But I am not sure how the whole problem was broken down to AX=B problem :/ Can someone point me to what algorithm is this solution based on? Btw, the code works like charm and transforms PS Move on to the right hand of Kinect!

```
private void CalculateTransformation()
{
Matrix moveMatrix;
Matrix kinectMatrix;
//Move Matrix of 50x4 and Kinect Matrix of 50x3
moveMatrix = Matrix.Zeros (samples_PSMove.Count, 4);
kinectMatrix = Matrix.Zeros (samples_Kinect.Count, 3);
for (int i = 1; i <= samples_PSMove.Count; i++) {
moveMatrix [i, 1] = new Complex (samples_PSMove [i - 1].x);
moveMatrix [i, 2] = new Complex (samples_PSMove [i - 1].y);
moveMatrix [i, 3] = new Complex (samples_PSMove [i - 1].z);
moveMatrix [i, 4] = new Complex (1.0f);
}
for (int i = 1; i <= samples_Kinect.Count; i++) {
kinectMatrix [i, 1] = new Complex (samples_Kinect [i - 1].x);
kinectMatrix [i, 2] = new Complex (samples_Kinect [i - 1].y);
kinectMatrix [i, 3] = new Complex (samples_Kinect [i - 1].z);
}
//perform a matrix solve Ax = B. We have to get transposes and inverses because moveMatrix isn't square(x,y,z,w)
//the solution is the same with (A^T)Ax = (A^T)B -> x = ((A^T)A)'(A^T)B
Matrix transformMatrixSolution = (moveMatrix.Transpose() * moveMatrix).Inverse() * moveMatrix.Transpose() * kinectMatrix;
// Matrix error = moveMatrix * transformMatrixSolution - kinectMatrix;
transformMatrixSolution = transformMatrixSolution.Transpose();
List<Vector3> orthogonalVectors = MathUtil.Orthonormalize(MathUtil.ExtractRotationVectors(MathUtil.MatrixToMatrix4x4(transformMatrixSolution)));
rotationMatrix = CreateRotationMatrix(orthogonalVectors);
//Debug.Log(rotationMatrix);
transformMatrix = MathUtil.MatrixToMatrix4x4(transformMatrixSolution);
UpdateFloorNormalAndDistance();
coordinateSystem.SetDeviceToRootTransforms(transformMatrix);
coordinateSystem.SaveTransformDataToXML(xmlFilename, RUISDevice.PS_Move, RUISDevice.Kinect_1);
coordinateSystem.SaveFloorData(xmlFilename, RUISDevice.Kinect_1, kinect1FloorNormal, kinect1DistanceFromFloor);
Quaternion rotationQuaternion = MathUtil.QuaternionFromMatrix(rotationMatrix);
Vector3 translate = new Vector3(transformMatrix[0, 3], transformMatrix[1, 3], transformMatrix[2, 3]);
updateDictionaries(coordinateSystem.RUISCalibrationResultsInVector3,
coordinateSystem.RUISCalibrationResultsInQuaternion,
coordinateSystem.RUISCalibrationResultsIn4x4Matrix,
translate, rotationQuaternion, transformMatrix,
RUISDevice.PS_Move, RUISDevice.Kinect_1);
coordinateSystem.RUISCalibrationResultsDistanceFromFloor[RUISDevice.Kinect_1] = kinect1DistanceFromFloor;
coordinateSystem.RUISCalibrationResultsFloorPitchRotation[RUISDevice.Kinect_1] = kinect1PitchRotation;
kinect1ModelObject.transform.rotation = kinect1PitchRotation;
kinect1ModelObject.transform.localPosition = new Vector3(0, kinect1DistanceFromFloor, 0);
psEyeModelObject.transform.position = coordinateSystem.ConvertLocation(Vector3.zero, RUISDevice.PS_Move);
psEyeModelObject.transform.rotation = coordinateSystem.ConvertRotation(Quaternion.identity, RUISDevice.PS_Move);
if(this.floorPlane)
this.floorPlane.transform.position = new Vector3(0, 0, 0);
}
```

**Answer** by kumarsmurthy
·
Aug 25, 2015 at 10:54 AM

I guess my question could be more simpler. I know that this is solving AX = B type of problem. Why has the author added one more row for PS move and not for Kinect? Is he using homogeneous coordinates? Can we do Ax=B when A is in homogeneous coordinates and B is in Cartesian?

**Answer** by masterchop
·
Aug 12, 2016 at 12:02 PM

OMG!!!!! i have been looking for this for a while, i cant believe i am looking at the code but i am in a similar situation, this code seemed to be for some sort of SDK, not the microsoft SDK right? or you think this will work with Kinectv2 with MS SDK?

i really need this to happen do you have more information?

### Your answer

### Welcome to Unity Answers

The best place to ask and answer questions about development with Unity.

To help users navigate the site we have posted a site navigation guide.

If you are a new user to Unity Answers, check out our FAQ for more information.

Make sure to check out our Knowledge Base for commonly asked Unity questions.

If you are a moderator, see our Moderator Guidelines page.

We are making improvements to UA, see the list of changes.

### Follow this Question

### Related Questions

How to move the camera depending on the score? 1 Answer

Transform an object based on another object's velocity? 0 Answers

Need help updating camera position to go above the player 2 Answers

Place object in front of me with given Translation Vector 0 Answers

UNET: Network Transform with skeleton? 3 Answers