عنوان مقاله [English]
Part of the studies of geomorphology is dedicated to the automated, semi-automated identification, segmentation and classification of landscapes and landforms at different scales. Each of the landscape classification systems, includes a number of smaller units or landforms. Some methods have been used to identify and recognize Landforms, while others have been dedicated to the landscape classifications. Landscape classification is applicable in a wide range of geomorphological studies such as, mapping geomorphological maps, zonation and environmental potential in the field of ecotourism, environmental exploitation and sustainable development, and also in the field of economic geography, natural hazard assessment and arranging the land use planning documents of the country and many other fields which is directly and indirectly applicable. In this study, an attempt has been made to present a new system in the landscape classification which be able to recognize and classify, landscapes of the terrain surface using, digital elevation models considering the ease and simplicity of the procedures. For this purpose, tangential, plan and profile curvature and slope extracted from digital elevation model, these three curvature combined together to get mean curvature and these factors including elevation, slope and mean curvature used to classify landscapes of the terrain. In the next step, each of the three components of the above-mentioned was divided into two parts based on 5 methods of thresholding, including: geometrical interval, quantile, natural breaks, standard deviation and weighted average. Finally, all three components have been coded and named with a special system and the area of the Iran, was classified into 8 landscapes units and the results were presented as a color map.
The Earth's surface can be represented as a mosaic of different form units, each of which has its own form-specificity and process. The classification of these landscapes into smaller units is of great importance and, of course, has many applications in many different disciplines. In many definitions of the geomorphology the study and identification of land form units has been emphasized. Part of the geomorphological study is dedicated to the automated and semi-automated systems of landscapes and landform classifications at different scales. Each of the landscape systems includes a number of smaller units which is called landforms. Some methods have been used to identify and recognize Landforms, and some others have implemented to detect and classify landscapes. Terrain landscape recognition and classifications are used in a wide range of geomorphological studies such as preparation of geomorphological maps, assessments and zonation of environmental capacities in the field of ecotourism, environmental exploitation and sustainable development, in the field of economic geography, environmental risk assessment, regulation the country's planning programs and many other sections, which are directly and indirectly applicable. In this study, it is tried to identify, recognize and classify terrain landscapes by using a new approach, based on the three main relief components of elevation, slope and curvature.
For classification of geomorphological landscapes in Iran, three morphological factors including: elevation, slope and mean curvature (combining plan, profile and tangential curvatures) implemented. As a result, these morphological components were used to define a new classification system. The elevation, slope and curvature of the terrain surface is obtained from digital surface model. All these factors are extracted from digital surface model with using moving window technique in the raster analysis environment. So a moving window of 3x3 in size implemented to calculate common slope algorithm in the field of geographical information system. In the next step, tangential, plan and profile curvature combined together to get mean curvature of the terrain surface whereas all these factors obtained from digital surface model in the same way. Then, each of these three form-components is divided into two classes. To determine the classification threshold (classes break) for all of the factors mentioned above, 5 thresholding methods including geometrical interval, natural breaks, quantile, standard deviation and weighted average had taken into account. From the combination of the three components of elevation, slope and curvature, each of which has two states, 8 units of landscapes were defined. Finally, 5 maps of Iran's geomorphological terrain landscapes, that each of them includes 8 landscape units, were implemented using Python programming and presented in the form of large-scale maps of Iran.
Results and discussion
Among the various methods used in the classification of terrain landscapes, geometric interval, quantile, natural break, and weighted average represent better results than the standard deviation method. In the first, third, and fourth standard deviation band thresholds, the results are not graphically appropriate, in contrast the second band of the standard deviation shows acceptable results. Natural breaks thresholding system, performs more sensitive in the higher values of the range whereas this point is visible in the output results. The quantile method offers similar results to the geometric interval, except that the geometric interval works more sensitive in the context of higher values in the variation range. Standard deviation methods generally do not show good performance in the classification of terrain landscape units, but in general, the second and third standard deviation bands provide better results. Finally, the results obtained from the weighted average method have acceptable performance and have a function among the methods of geometrical interval and natural breaks.
In the present study, the landscapes of the earth were classified with a new approach. In this system, some major morphological factors had taken into account. To determine the appreciate threshold, geostatistical thresholding methods used and the results of these methods, investigated. The classification of terrain landscapes with the combined system presented in this study, which is based on a completely morphological concepts, helps to improve the methods of classification of terrain landscapes in geomorphology and the results would be suitable to use in the field of environmental and land use planning. Terrain landscape classification, from the morphological point of view, is applicable in the field of potential ability of the environment, echological potentials of the landscapes and so on and could effectively increse the function of the geomorphology in this scope.