These keywords were added by machine and not by the authors. We are still far from being able to develop a common framework to successfully identify a variety of features in different landscapes and to generalize and automate the classification process. However, these methods require extensive training and human supervision. Newer approaches such as fuzzy classification, artificial neural network, and object-based classification have been developed and successfully applied (Definiens, 2004 Benz et al., 2004). As a result, new and sophisticated classification methods designed to improve the classification process continue to appear in the literature (e.g., Jensen, 2005 Gong, 2006). Different land cover/land use classification methods may be needed for different problems under different environmental conditions, making generalization and hence automation of the image classification process across time and space extremely difficult. Unfortunately, there are numerous factors (e.g., image resolution and atmospheric condition) that could affect the effectiveness and accuracy of the classification algorithms. Mapping land cover/land use accurately and efficiently via remote sensing requires good image classification methods. Knowledge of land cover/land use and its change is also critical to effective planning and management of natural resources. More importantly, land cover/land use pattern and its change reflect the underlying natural and/or social processes, thus providing essential information for modeling and understanding many different phenomena on the Earth. Land cover/land use has been used extensively to derive a number of biophysical variables, such as vegetation index, biomass, and carbon content (see other chapters).
Mapping and identifying land cover/land use and its change is the most important, as well as the most widely researched, topic in remote sensing.
0 kommentar(er)
