To make the overlays from which the 1969 land-use map was constructed, the topographic map was scanned at 200 dots per inch and then imported into the IDRISI program where it was georegistered using the latitude and longitude marks at the margins. The map was then reprojected to Universal Transverse Mercator (UTM47N), using the parameters of the ASTER and DEM maps already prepared. The map was exported in graphics format (BMP) for processing outside IDRISI. So long as the pixel dimensions of the exported image are not changed during processing in the graphics program (Photoshop), the image can be re-imported into IDRISI and easily georeferenced. (This is done by swapping the metadata (RDC file) of the image before export for the RDC file of the image after re-importing. If the maximum and minimum values of the re-imported image are different from those values in the exported image, the metadata of the re-imported image will have to be changed manually.)
Using the "Magic Wand" tool in Photoshop, the outline of an area can be marked and filled with a color that does not appear elsewhere in the image. For example, all marshland can be changed to yellow. Using one of several methods, all other colors and black are then removed from the image, leaving only marshland yellow with a white background. Marshland can then be changed to black and the image imported into IDRISI, where the routine "RECLASS" is used to change values to one and zero.
The resulting image of ones and zeros is binary and is called "Boolean" because the values "1" and "0" can be used in a logical sense to indicate the presence (TRUE) or absence (FALSE) of a feature. For example, two images can be multiplied so that the resulting image shows the areas where both features are present as black (TRUE). Adding or subtracting images produces different results. For example, image A (below) minus image F (below) produces a map of land (black) and sea (white) with the airport zero (white).
The binary concept can be extended beyond Boolean. I processed all the binary images by first multiplying them as follows: image B times 2, C times 4, D times 8, E times 16, and F times 32. I then made an image that was the sum of all six images.
The reason for this operation is the peculiar nature of binary numbers representing the decimal numbers in the series 1, 2, 4 to 32. Each binary number has a single digit with a value of one and the rest zero. For example, the number eight as a binary number is 000001000. Thirty-two is 000100000. Their sum is forty, 000101000. Since the original numbers are not confounded by this process, the original numbers can be recovered by inspection. Similarly, adding all of the images A to F produces a resulting image with numbers that reveal their original components on inspection.
After multiplication, then adding images A to F below and a seventh for buildings (not shown), I inspected the numbers in the resulting image. Applying the RECLASS routine of IDRISI, I used these numbers to group the pixels within the eight classes shown in the 1969 land-use map (Figure 7). The class, "other land" is a residual land category. Only seven images were needed for eight categories because two categories, sea and land, are included in image A. Colors were assigned using the palette facility of the IDRISI Symbol Workshop.
Images G and H indicating elevation were converted to polygons (vector images) and superimposed on the land-use class map to assist with interpretation. (Image G was generalized by removing small areas before vectorizing.) To show elevation, patterns and outlines were assigned using the polygon section of the IDRISI Symbol Workshop.
