I’m attempting to transform TrueDepth depth maps into level clouds. Is there any solution to rework TrueDepth digital camera depth map into an actual dimension level cloud utilizing supplied by Apple intrinsics? Different questions do not assist. That is how I extract parameters:
if let calibrationData = depthData.cameraCalibrationData {
let intrinsicMatrix = calibrationData.intrinsicMatrix
let extrinsicMatrix = calibrationData.extrinsicMatrix
let referenceDimensions = calibrationData.intrinsicMatrixReferenceDimensions
let intrinsicReferenceDimensionWidth = referenceDimensions.width
}
the place depthData: AVDepthData
My transformation code is right here (with IPhone 12 intrinsics and reference dimensions divided by depth dimensions):
def depth_to_point_cloud(depth, intrinsics=[[2742.1404 / (4032 / 640), 0, 2015.808 / (4032 / 640)], [0, 2742.1404 / (4032 / 640), 1513.418 / (4032 / 640)], [0, 0, 1] ]):
peak, width = depth.form
x, y = np.meshgrid(np.arange(width), np.arange(peak))
xrw = (x - intrinsics[0][2]) * depth / intrinsics[0][0]
yrw = (y - intrinsics[1][2]) * depth / intrinsics[1][1]
xyzw = np.stack([xrw, yrw, depth], axis=2)
return xyzw.reshape(-1, 3)
The outcome could be very stretched and doesn’t seem like some extent cloud with actual distances. What am I lacking? Any assist, questions, strategies, hyperlinks, guides, literature could be appreciated!
