A comparitive study of land classification using remotely sensed data

Land classification is the process of surveying countryside characteristics such as land form, soils and ecosystem. They may be aimed firstly at assessing the agricultural or the forestry potential, or secondly they may be a simple categorisation and mapping of specific characteristics. The land classification imparts knowledge about land use and land cover has become increasingly important as the country plans to overcome the problems of haphazard, unplanned development, decreasing environmental quality, loss of good agricultural lands, destruction of cultivation lands, important water catchments, and wildlife habitat. There are many types of land classification algorithms available in remote sensing method such as Minimum Distance, Maximum Likelihood, Support vector machines, k-NN and Multi-Label Classification (MLC). A comparative analysis of land cover for all classifiers was done based on three factors 1) overall classification 2) accuracy 3) performance in the heterogeneous area. The survey concludes that the Multi-Label method classifier will produce better results.