Building a Transformation Grid
The process has three phases:
Calc Displacements
Select Points/advanced/NTv2 Calc Displacements and enter the reference layer and the layer to adjust. The program will
match common points, compute the error vector between them and draw a contour map of the displacements in the X direction.
The search routine for common points starts by matching parcels by name, then by matching the points within the matched
parcels. A “robust estimator” technique is used to select or reject points according to the shape differences between the
two matched parcels. The rejection criteria (the rejection factor) can be changed to allow for different types of data.
The contour interval will be 200 mm for minor contours and one metre for major contours.
The contour map is a graphic way of displaying the trends in the differences in the two layers and each pair of points are
also linked with a blue line. The contours are generated by first building a triangular mesh over all of the matched points.
This layer can be turned on to show the matched points if required .If there is any discontinuity in the data such as a
mis-matched point, it will show up as a “whirl-pool” in the contours. These can be changed using the edit option.
Edit Displacements
This option allows the difference vectors to be deleted or and new vectors inserted as needed .
To delete a vector, select the point on the target layer with the mouse. Its point number will be displayed in the “Point 1 “ box.
Select the delete button in the edit menu and the vector will be deleted.
To add a vector, first select the point on the layer to be adjusted, then (Point 1), then select the point on the
Reference Layer (Point 2) and select the “Add” button to add the vector. To view the effect of the corrections use the
“contour dX “ or “Contour dY” buttons to display the relevent contour map of displacements.
When the correction vectors are satisfactory, a correction file can be exported.
There are three types of correction files as follows:
Write NTv2 File.
An NTv2 transformation grid file contains coordinate shift values at nominated grid nodes in a format known as NTv2
(National Transformation Version 2). This format was developed by the Geodetic Survey Division of Geomatics Canada,
to implement the transformation of coordinates between NAD27 and NAD83 in Canada.
The format of the grid file (NTv2) is described in the following web page:
http://www.geom.unimelb.edu.au/gda94/UsersGuide.pdf
The format has been adopted by many countries (including Australia) to provide a means of transforming coordinates between
different geodetic datums. Most data base systems have software to apply this transformation procedure.
A Grid Shift file contains one or more rectangular “sub grids". Each sub grid consists of nodes spaced at regular
intervals of latitude and longitude. Each node has a shift value for both latitude and longitude components, as well as a
value for the accuracy of each shift.
In its simplest form, a Grid Shift file defines a single sub grid, however it can contain more than one sub grid.
If multiple sub grids are defined, the sub grids do not have to be contiguous. Sub grids of different densities can also
overlap each other if certain rules are adhered to.
The program can build a NTv2 grid to reflect the differences between two cadastral mapping systems so that the associated
layers in a data base can me moved to reflect the change in the cadastral layer.
The program will write a NTv2 sub grid for an area selected in a view port where the correction vectors have been calculated.
The mesh size can be set and for urban areas, a mesh size of 50 metres is recommended. A larger mesh size may be more
appropriate for rural areas. Note that if you halve the mesh size, there are four times the number of elements, and if you
quarter the mesh size, there are sixteen times the number of elements.
Create a viewport using the viewport option, select the viewport and the output file name and the file will be written.
The viewport borders decide the total mesh size.
To test the file, restart the program with the layer to be adjusted. Apply the correction using Points/Advanced/Datum Change,
then read in the reference layer. If the process has been carried out correctly. The two cadastral patterns will now be
exactly superimposed on top of each other.
Write ESRI Correction File.
THe ESRI correction file has our values per line with a TAB separating each value.
One line is written for each matching point and the values are the x and y of the reference layer and the x and y of the adjust layer.
A few lines of a file is below
497794.746301 6914931.843439 497796.471300 6914928.457400
497786.951439 6914887.019594 497788.530500 6914883.626500
499292.461352 6914739.170131 499289.958600 6914737.036400
499305.784571 6914732.946461 499303.325100 6914730.807600
499317.177939 6914723.665787 499314.762100 6914721.509500
Write the FME Correction File.
THe FME correction file has five values per line with a comma separating each value.
One line is written for each matching point and the values are the x and y of the adjust layer, x and y of the reference layer and the length of the vector between the two points.
A few lines of a file is below
497796.471,6914928.457,497794.746,6914931.843,3.800
497788.530,6914883.627,497786.951,6914887.020,3.743
499289.959,6914737.036,499292.461,6914739.170,3.289
499303.325,6914730.808,499305.785,6914732.946,3.259
499314.762,6914721.509,499317.178,6914723.666,3.238
Testing the Transform File.
To test the transform, restart the program and read in the file to be adjusted.
Apply the correction using Points/Transforms/User Transform Job.
Then append the reference layer
If the process has been carried out correctly, the two cadastral patterns will now be
superimposed on top of each other.