استخراج و شناسایی لندفرم های یخچالی با استفاده از روش شی‌گرا (مطالعه موردی سیرک های یخچالی سبلان)

Introduction
Semi-automated extraction of selected landform types from land-surface models such as  digital elevation models (DEMs), curvatures, and slope gradients is of particular interest in  geomorphology, hydrology and related fields(Eisank et al,2010:1). Geomorphometry is the science of quantitative land-surface analysis (Pike, 1995, 2000a; Rasemann et al., 2004). The morphometry of landforms perse, by or without the use of digital data, is more correctly considered part of quantitative geomorphology (Thorn, 1988; Scheidegger, 1991 Leopold et al., 1995; Rhoads and Thorn, 1996).
Evans (1972) was the first person that divided Geomorphometry in two branches the general and special Geomorphometry, in special Geomorphometry More is be discussed landforms and geomorphic processes such as glaciers and runoff in geomorphometry terrain segmentation is an approach for structuring terrain data such as digital elevation model(DEM) via derived land-surface into spatially discrete plan-view areas known as terrain units(strobl,2008).The shape of terrain, i.e. landforms, influence the flow of surface water, transport of sediments, soil production, and determines climate on local and regional scales, furthermore, natural phenomena like vegetation are directly influenced by landform patterns and their relative position across the landscape Landforms although have different meaning to different disciplines, they reveal common physiological and morphological characteristics of terrain which may guide through understanding past and present processes acting on terrain and provide necessary information to related disciplines about land characteristics and potentials(Gerçek, 2010:9). In this study, using MRS algorithms and Ecognition software have been studied glacial landforms the northern slopes of Sabalan.
 Methodology
    In this study, the DEM and derivative of its, have been used in order to extract glacial landforms. To prepare the digital elevation model layer have been used of topographic maps with 1: 25000 scale. Profile curvature, plan curvature and mean curvature are three important layers used in this study.The semi-automated methods refer to the automatic procedures of extracting landform based-process. This is mainly relying on unsupervised isodata classification, pixel-based clasification (supervised /subpixel classifier based on training material), analysis of digital elevation models (DEM), algorithms, hydrological modelling and object oriented analysis (Nabil and Moawad, 2014:42). In this study object-oriented methods and Ecognition software is used for classification and extraction of landforms.
 Results and discussion
In this study, we are looking for are extracted cirques the northern slopes of Sabalan. This area because of the high altitude receives high humidity this precipitation is mainly in the form of snow and is appropriate for the formation of glacier and for cirques. A cirque is one of the most prominent forms of glacial erosion (Embleton   and Hamann, 1988). According to Evans (2004: 154) cirques are “hollows formed at glacial sources in mountains and partly enclosed by steep, arcuate slopes    (headwalls)”. In general cirques exhibit concave shape in both plan and profile   direction, with a size ranging from hundreds of metres to a few kilometers (Glasser and Bennett, 2004). They are typically composed of a cirque crest, a steep headwall, and a gentler cirque floor often filled by a cirque lake. For extraction of circus having profile curvature, plan curvature and mean curvature layers is necessary. With respect to this aforementioned layers were obtained in the SAGA software.
Scale parameter is a crucial threshold that determines the maximum allowed heterogeneity for   segmentation, which   has   a   direct   influence   on   the   size   of   the   objects   to   be obtained. Scale parameter is recognized after a trial and error process (Gerçek, 2010:115).  In this study ultimately scale of 36 was selected for the segmentation.
Extracting ridge line position is very important in determining and identifying the the cirque. According to the morphometry of this landforms for classification the ridge were used convex shapes in layers.
In general, the convex shape of the ridge is determined by the positive values in mean and plan curvature. The Ecognition software and MRS algorithm was used for the execution of the model and Segmentation of the layers. Ridge layer was obtained by defining morphometriy characteristics of the ridge and comparing the results with field data. Finally, according to the semantic model of research, the relevant rules defined in Ecognition software and was extracted circus position.
 Conclusion
In this study glacial landforms (circus) were extracted with semi-automatic approach and by using object-oriented approach in northern slopes of Sabalan. The Result of research were presented in the form of a map which shows the location circuses.