[ad_1]
I’am creating an app that permits person to scan one thing with TrueDepth Digital camera utilizing AVCaptureDataOutputSynchronizer.
I’ve already carried out scanning by single depth body. What I would like now could be to create single PointCloud with a number of depth frames.
I assume that I must multiply all factors from every body to some transformation matrix (appears to be like prefer it’s the ExtrinsicMatrix) to get appropriate factors coordinates by assuming appropriate digicam pose (translation and rotation).
I can obtain this extrinsicMatrix with cameraCalibrationData.extrinsicMatrix however it’s at all times an id matrix. Apple docs says
If the rotation matrix is an id matrix, then this digicam is the reference digicam.
However I assume I’d have just one digicam in app. Or is it potential to create digicam that describes the primary body and course of all different frames evaluating each cameras? As a result of apple have this
class func extrinsicMatrix(from fromDevice: AVCaptureDevice, to toDevice: AVCaptureDevice) -> Knowledge?
So essential query is: how can I get a matrix that comprises digicam place in world area in AVCapture
[ad_2]
