Changing Coordinate Systems in Idrisi

Changing Coordinate Systems in Idrisi

The Resample Module

GIS EA 2003

Often in GIS you have map data developed without a coordinate system. We will frequently use satellite imagery in this class, and raw satellite data does not have a coordinate system. Other examples are maps without a coordinate system like those in many of the government lake reports and the 2 volumes of Lakes of Washington. One way to rectify these data is to use a polynomial transformation of the unknown coordinates to known coordinates. The module Resample does this in Idrisi. It is important enough I have copied the help menu below.

RESAMPLE

Menu Location:

Analysis/Image Processing/Restoration

Reformat

RESAMPLE registers the data in one grid system to a different grid system covering the same area. The process uses polynomial equations to establish a rubber sheet transformation, as if one of the grids were placed on rubber and warped to make it correspond to the other. The actual process is one in which a new grid is constructed and a set of polynomial equations is developed to describe the spatial mapping of data from the old grid into the new one. The new grid is then filled with data values by resampling the old grid and estimating, if necessary, the new value. Resampling options include a nearest neighbor option in which the new grid value is the same as that in the nearest cell in the old grid, and a bilinear interpolation option in which the new value is a distance weighted average of the four nearest neighbors in the old grid. RESAMPLE is used for a variety of operations, including:

- registering remotely sensed imagery to a grid referencing system;

- registering maps with different reference systems over small areas (larger areas require a projection change) when one or both reference systems are not supported by PROJECT;

- undertaking minor changes in projection when one or both reference systems are not supported by PROJECT;

- making non-integer changes in the resolution of an image.

Note that when both the original and new reference systems are supported, the PROJECT module will yield a more precise re-mapping and should be used.

In this exercise you will take a vector map of the campus buildings where the coordinates are in the TESC Campus Grid," a local coordinate system used for campus surveys and drawings. These data are more useful when transformed to a more widely used system like Washington SPCS or UTM. The original campus survey provided coordinates for several points in both Campus Grid and Washington SPCS, south zone, NAD27. See table 1 for these data.

Table 1. Campus and State Plane Coordinate System control points. All in feet.

Campus / Campus / SPCS NAD27 / SPCS NAD27
E / N / E / N
4770.94 / 14893.56 / 1379600.95 / 645409.93
7100.00 / 15000.00 / 1381932.18 / 645391.86
7100.00 / 13300.00 / 1381841.32 / 643694.29
10000.00 / 15000.00 / 1384828.03 / 645236.87
10000.00 / 12364.06 / 1384687.15 / 642604.70
10093.21 / 9698.92 / 1384637.79 / 639938.40

Carefully read about the operation of RESAMPLE in Idrisi help. The correspondence file referred to must have the extension .cor (unfortunately corrected Trimble GPS files also use this extension).

The vector data are in the folder Wk-2 resample in the program share. Copy the folder, as usual, to C:\temp. You must make the correspondence file, space delimited as instructed in the help menu. Using Idrisi's Edit module might be the best way to do this.

For the resample use a linear transformation and be sure to enter the NEW coordinate points when called for in the right units and with the correct extents. For convenience use the extents (in WA SPCS NAD27) in Table 2 for the campus buildings.

Table 2. Extents in feet for the "resampled" building vector image.

Min X / 1380000
Max X / 1389500
Min Y / 641000
Max Y / 650000

Run the Resample module and view the results. You will see the new X and Y coordinates on the screen as you pass the arrow over the buildings.

Note the Resample module can do a transformation that stretches and pinches the image to get a the best fit of the points. For this reason the process is often called "rubber sheeting." Try to understand what is happening in general terms. It is an extremely useful spatial tool.