The use of Landsat TM data to detect land use and land cover changes has been generally a success.
The digital image classification coupled with GIS has demonstrated its ability to provide comprehensive
information on the direction, nature, rate, and location of land use and land cover changes as a result of rapid industrialization and urbanization. However, the issue of class uncertainties in image classification has not been examined in this paper. Although the land use and land cover maps have a reasonably high overall accuracy, the accuracy of different classes varies. The horticulture farms class was confused with several other classes due to their diversity. This confusion has hindered the obtainment of accurate TPs. Moreover.