Abstract

The objective of this research focuses on comparing Landsat TM and IRS data and determining if similar classification can be achieved from datasets for certain land cover types. Supervised classification was performed using information from a combination of digital aerial photographs, a priori knowledge of the study site by the authors and existing Land Use Land Cover (LULC) maps. The “upland forest,” “open water,” “tree crops” and “palmetto prairie” categories show strong agreement in terms of percentage of LULC found in both Landsat TM+ and IRS classified images. Conversely, the “open land,” “cropland and pastureland” and “wetlands” categories display differences based on the land cover area. Based on the overall classification accuracies similar results were produced for both TM and IRS data of 86.3% and 88.4% respectively. On the other hand, certain LULC categories did not perform so well, such as the golf course. Temporal resolution between the TM and IRS images was six weeks, and this was considered a factor in the confusion between LULC category discrimination. This study showed that using Landsat TM and IRS in same study provide promising results for LULC studies