Overview

The following sections describe the process of satellite image classification, whereby image pixels are automatically assigned (by the classification software) to one of a number of output classes. Two methods of image classification were utilised, supervised and unsupervised. Supervised classification, as the name suggests, involves the user defining, by delineating training sites, the output classes to which pixels will be assigned. On the other hand, unsupervised classification or image clustering requires no pre-classification input from the user and pixels are split into a number of groups (to be specified by the user), based on their spectral similarity. Supervised classification formed the core of the classification work, while an unsupervised classification was performed to assist in the selection of training sites.

Table 4.2c lists the classes that were mapped through supervised classification. These classes represent one or more habitat types that were spectrally inseparable. An example of spectral inseparability are the habitat types Bare Rock/Soil and Landfill, which are potentially composed of similar surface features (exposed soil).

(8) Since the contact print has a scale of 1:40,000 (Section 4.1.2), each inch on the print represents 40,000 * 0.0254 metres on the ground, and since there are 600 dots per inch, each dot therefore represents 40,000 * 0.0254 / 600, or 1.69 metres.

From Table 4.2c, it is seen that the habitat type Mixed Shrubland was subdivided into two classes Low Mixed Shrub and Tall Mixed Shrub after initial classifications resulted in low accuracy for this habitat type. This is because in the initial classification, there was a tendency for low shrub habitats to be mis-classified as Shrubby Grassland and tall shrub as Forest. Low Mixed Shrub was defined as those below one metre. At the end of the classification, these two classes were merged to become the Mixed Shrubland habitat.