uvTo3DInCameraSpace
Using PICO's depth sensor and stereo view RGB images, estimate the 3D coordinates of 2D points on camera output.
Parameters
the 2D coordinates of points, corresponding to one left-eye camera image. The tensor must be a point2 array, (i.e., created with Tensor.Point2ArrayInitInfo) of size N. The data type must be DataType.INT32, where (0, 0) is top-left corner of the camera image, and the (H, W) is the bottom-right corner, where H and W are the resolutions of the session, where the SpatialMLSession is initialized.
the camera timestamp of the left-eye image to which the uv is corresponding. It must be a timestamp tensor output from rectifiedVSTAccess.
the camera intrinsic matrix of the left-eye image to which the uv is corresponding. It must be a multi-dimensional tensor of 1 channel, dimensions = 3x3 and its data type should be 32-/64-bit float. You can directly use the output from rectifiedVSTAccess.
the raw left-eye camera image from rectifiedVSTAccess. It must be the output associated with the timestamp. It is used to enhance the depth sensor's output using parallax.
the raw right-eye camera image from rectifiedVSTAccess. It must be the output associated with the timestamp. It is used to enhance the depth sensor's output using parallax.
the required result from this operation, to store the estimated 3D coordinates corresponding to the 2D points in uv. It must be a point3 array, created with Tensor.Point3ArrayInitInfo of size N, where N must match that of uv, or a multi-dimensional tensor with dimensions = Nx3 and of non-pixel datatype or RED-channel- pixel datatype. Alternatively, it can also be a multi-dimensional tensor with dimensions = Nx1 or 1xN, and of 3-channel pixel datatype (R, G and B channels will store the X, Y and Z coordinates of a point).
Throws
If the usage is not allowed by the SpatialML run-time Framework, or if the framework encounters internal error and cannot perform the requested behavior.