Helmert transform

The Helmert transformation changes coordinates from one reference frame to anoether by means of 3-, 4-and 7-parameter shifts, or one of their 6-, 8- and 14-parameter kinematic counterparts.

Input type

Cartesian coordinates (spatial), decimalyears (temporal).

Output type

Cartesian coordinates (spatial), decimalyears (temporal).

Options

x

Translation of the x-axis [m]. Optional.

y

Translation of the y-axis [m]. Optional.

z

Translation of the z-axis. [m]. Optional.

s

Scale factor [ppm]. Optional.

rx

X-axis rotation in the 3D Helmert [arc seconds]. Optional.

ry

Y-axis rotatoin in the 3D Helmert [arc seconds]. Optional.

rz

Z-axis rotation in the 3D Helmert [arc seconds]. Optional.

theta

Rotation angle in the 2D Helmert. [arc seconds]. Optional.

dx

Translation rate of the x-axis. [m/year]. Optional.

dy

Translation rate of the y-axis. [m/year]. Optional.

dz

Translation rate of the z-axis. [m/year]. Optional.

ds

Scale rate factor [ppm/year]. Optional.

drx

Rotation rate of the x-axis [arc seconds/year]. Optional.

dry

Rotation rate of the y-axis [arc seconds/year]. Optional.

drz

Rotation rate of the y-axis [arc seconds/year]. Optional.

t_epoch

Central epoch of transformation. [decimalyear]. Only used in spatiotemporal transformations. Optional.

t_obs

Observation time of coordinate(s). Mostly useful in 2D and 3D transformations. [decimalyear]. Optional.

exact

Use exact transformation equations. Optional.

transpose

Transpose rotation matrix. Optional.

The Helmert transform, in all it’s various incarnations, is used to perform reference frame shifts. The transformation operates in cartesian space. It can be used to transform planar coordinates from one datum to another, transform 3D cartesian coordinates from one static reference frame to another or it can be used to do fully kinematic transformations from global reference frames to local static frames.

All of the parameters described in the table above are marked as optional. This is true as long as at least one parameter is defined in the setup of the transformation. The behaviour of the transformation depends on which parameters are used in the setup. For instance, if a rate of change paramater is specified a kinematic version of the transformation is used.

The kinematic transformations require an observation time of the coordinate, as well as a central epoch for the transformation. The latter is usually documented alongside the rest of the transformation parameters for a given transformation. The central eopch is controlled with the parameter t_epoch. The observation time can either by stated as part of the coordinate when using PROJ’s 4D-functionality or it can be controlled in the transformation setup by the parameter t_obs. When t_obs is specified, all transformed coordinates are treated as if they have the same observation time.

Examples

Transforming coordinates from NAD72 to NAD83 using the 4 parameter 2D Helmert:

proj=helmert ellps=GRS80 x=-9597.3572 y=.6112
             s=0.304794780637 theta=-1.244048

Simplified transformations from ITRF2008/IGS08 to ETRS89 using 7 parameters:

proj=helmert ellps=GRS80 x=0.67678    y=0.65495   z=-0.52827
            rx=-0.022742 ry=0.012667 rz=0.022704  s=-0.01070

Transformation from ITRF2000@2017.0 to ITRF93@2017.0 using 15 parameters:

proj=helmert ellps=GRS80
     x=0.0127     y=0.0065     z=-0.0209  s=0.00195
     dx=-0.0029   dy=-0.0002   dz=-0.0006 ds=0.00001
     rx=-0.00039  ry=0.00080   rz=-0.00114
     drx=-0.00011 dry=-0.00019 drz=0.00007
     epoch=1988.0 tobs=2017.0    transpose

Mathematical description

In the notation used below, \(\hat{P}\) is the rate of change of a given transformation parameter \(P\). \(\dot{P}\) is the kinematically adjusted version of \(P\), described by

(1)\[\dot{P}= P + \hat{P}\left(t - t_{central}\right)\]

where \(t\) is the observation time of the coordinate and \(t_{central}\) is the central epoch of the transformation. Equation (1) can be used to propagate all transformation parameters in time.

Superscripts of vectors denote the reference frame the coordinates in the vector belong to.

2D Helmert

The simplest version of the Helmert transform is the 2D case. In the 2-dimensional case only the horizontal coordinates are changed. The coordinates can be translated, rotated and scale. Translation is controlled with the x and y parameters. The rotation is determined by theta and the scale is controlled with the s paramaters.

Note

The scaling parameter s is unitless for the 2D Helmert, as opposed to the 3D version where the scaling parameter is given in units of ppm.

Mathematically the 2D Helmert is described as:

(2)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ \end{bmatrix}^B = \begin{bmatrix} T_x \\ T_y \\ \end{bmatrix} + s \begin{bmatrix} \hphantom{-}\cos \theta & \sin \theta \\ -\sin \theta & \cos \theta \\ \end{bmatrix} \begin{bmatrix} X \\ Y \\ \end{bmatrix}^A \end{align}\end{split}\]

(2) can be extended to a time-varying kinematic version by adjusting the parameters with (1) to (2), which yields the kinematic 2D Helmert transform:

(3)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ \end{bmatrix}^B = \begin{bmatrix} \dot{T_x} \\ \dot{T_y} \\ \end{bmatrix} + s(t) \begin{bmatrix} \hphantom{-}\cos \dot{\theta} & \sin \dot{\theta} \\ -\sin\ \dot{\theta} & \cos \dot{\theta} \\ \end{bmatrix} \begin{bmatrix} X \\ Y \\ \end{bmatrix}^A \end{align}\end{split}\]

All parameters in (3) are determined by the use of (1), which applies the rate of change to each individual parameter for a given timespan between \(t\) and \(t_{central}\).

3D Helmert

The general form of the 3D Helmert is

(4)\[\begin{align} V^B = T + \left(1 + s \times 10^{-6}\right) \mathbf{R} V^A \end{align}\]

Where \(T\) is a vector consisting of the three translation parameters, \(s\) is the scaling factor and \(\mathbf{R}\) is a rotation matrix. \(V^A\) and \(V^B\) are coordinate vectors, with \(V^A\) being the input coordinate and \(V^B\) is the output coordinate.

The rotation matrix is composed of three rotation matrices, one for each axis:

\[\begin{split}\begin{align} \mathbf{R}_X &= \begin{bmatrix} 1 & 0 & 0\\ 0 & \cos\phi & -\sin\phi\\ 0 & \sin\phi & \cos\phi \end{bmatrix} \end{align}\end{split}\]
\[\begin{split}\begin{align} \mathbf{R}_Y &= \begin{bmatrix} \cos\theta & 0 & \sin\theta\\ 0 & 1 & 0\\ -\sin\theta & 0 & \cos\theta \end{bmatrix} \end{align}\end{split}\]
\[\begin{split}\begin{align} \mathbf{R}_Z &= \begin{bmatrix} \cos\psi & -\sin\psi & 0\\ \sin\psi & \cos\psi & 0\\ 0 & 0 & 1 \end{bmatrix} \end{align}\end{split}\]

The three rotation matrices can be combined in one:

\[\begin{align} \mathbf{R} = \mathbf{R_X} \mathbf{R_Y} \mathbf{R_Y} \end{align}\]

For \(\mathbf{R}\), this yields:

(5)\[\begin{split}\begin{bmatrix} \cos\theta \cos\psi & -\cos\phi \sin\psi + \sin\phi \sin\theta \cos\psi & \sin\phi \sin\psi + \cos\phi \sin\theta \cos\psi \\ \cos\theta\sin\psi & \cos\phi \cos\psi + \sin\phi \sin\theta \sin\psi & - \sin\phi \cos\psi + \cos\phi \sin\theta \sin\psi \\ -\sin\theta & \sin\phi \cos\theta & \cos\phi \cos\theta \\ \end{bmatrix}\end{split}\]

Using the small angle approxition the rotation matrix can be simplified to

(6)\[\begin{split}\begin{align} \mathbf{R} = \begin{bmatrix} 1 & -R_z & R_y \\ Rz & 1 & -R_x \\ -Ry & R_x & 1 \\ \end{bmatrix} \end{align}\end{split}\]

Which allow us to express the most common version of the Helmert transform, using the approximated rotation matrix:

(7)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^B = \begin{bmatrix} T_x \\ T_y \\ T_z \\ \end{bmatrix} + \left(1 + s \times 10^{-6}\right) \begin{bmatrix} 1 & -R_z & R_y \\ Rz & 1 & -R_x \\ -Ry & R_x & 1 \\ \end{bmatrix} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^A \end{align}\end{split}\]

If the rotation matrix is transposed the transformation is effectively reversed. This is cause for some confusion since there is no correct way of defining the rotation matrix. Two conventions exists and they seem to be equally popular. In PROJ the rotation matrix can be transposed by adding the transpose flag in the transformation setup.

Applying (1) we get the kinematic version of the approximated 3D Helmert:

(8)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^B = \begin{bmatrix} \dot{T_x} \\ \dot{T_y} \\ \dot{T_z} \\ \end{bmatrix} + \left(1 + \dot{s} \times 10^{-6}\right) \begin{bmatrix} 1 & -\dot{R_z} & \dot{R_y} \\ \dot{R_z} & 1 & -\dot{R_x} \\ -\dot{R_y} & \dot{R_x} & 1 \\ \end{bmatrix} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^A \end{align}\end{split}\]

The Helmert transformation can be applied without using the rotation parameters, in which case it becomes a simlpe translation of the origin of the coordinate system. When using the Helmert in this version equation (4) simplies to:

(9)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^B = \begin{bmatrix} T_x \\ T_y \\ T_z \\ \end{bmatrix} + \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^A \end{align}\end{split}\]

That after application of (1) has the following kinematic countpart:

(10)\[\begin{split}\begin{align} \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^B = \begin{bmatrix} \dot{T_x} \\ \dot{T_y} \\ \dot{T_z} \\ \end{bmatrix} + \begin{bmatrix} X \\ Y \\ Z \\ \end{bmatrix}^A \end{align}\end{split}\]