Using a sophisticated orbit-attitude model which accurately replicates true viewing parameters, the satellite images are geometrically corrected to mapping co-ordinates. A digital elevation model derived from the NTDB data is used to compensate for terrain effects and thus produce a true ortho-image.

Each pixel on the panchromatic SPOT image represents 10 metres on the ground. On the Landsat colour image each pixel represents 30 metres on the ground. By splitting the colour TM data into its intensity and colour components and replacing the intensity component with the SPOT data, the two images are merged to create a colour image with 10 metre pixels. The resulting image achieves the best of both worlds.

The red, green and blue imagery is transformed on an electronic prepublishing system into cyan, magenta, yellow and black, the four primary colours used in printing, and the image is colour corrected. By creating a mask from the NTDB data, it is possible to correct the colour of water covered areas separately from land areas to produce a superior looking product.

At this stage the NTDB vector data overlay is incorporated into the imagery. Now it is possible to select any combination of map features that are to be included - such as roads, contours and hydrography. The four colour separations are then plotted on film using a high resolution laser plotter. Printing plates can be produced directly from these separations, because they are printed as halftone images, that is, each colour made up of dot patterns of the four primary colours. Since the processing is end-to end digital, any changes can be incorporated quickly and efficiently. Lessons learned while processing the data and feedback received from the user community contribute to ongoing refinements so that each new map is an improvement over the last.