When land cover classification is applied to terrain that is not flat, inaccurate classification results are produced. Terrain variations affect not only the position of a target on the Earth's surface, but also the brightness of the radar return. Without treatment, the radiometric biases caused by terrain variations are introduced into the coherency and covariance matrices. It is often seen that the classification result mimic the radiometry rather than the actual land cover. This output removes the radiometric variability associated with topography using the terrain flattening method proposed by Small [1] while leaving the radiometric variability associated with land cover.