سال انتشار: ۱۳۹۲
محل انتشار: کنفرانس بین المللی عمران، معماری و توسعه پایدار شهری
تعداد صفحات: ۹
s Talebi – Civil Engineering Faculty, Tafresh State University, Tafresh, Iran
p Pahlavani – Center of Excellence in Geomatic Eng. in Disaster Management, Dept. of Surveying and Geomatic Eng., College of Eng., University of Tehran, Tehran, Iran
Detection and extraction of objects in complex city scenes are challenging matters in many fields which are strongly in associate with an urban sustainable development. On this basis, detection of trees as city lungs plays an important role to get urban planners goals. In this paper, trees have been detected automatically in complex city scenes using an adaptive neuro-fuzzy inference system (ANFIS) with subtractive clustering based FIS. In this regard, the proposed ANFIS was trained by a hybrid learning algorithm. Accordingly, input data are features that are extracted from aerial images and LiDAR data. There are three test areas to check and evaluate efficiency of the proposed method. Evaluating the proposed methodology over the three different areas of Vaihingen in Germany achieved acceptable results in detecting trees in complex city scenes, in which in Area1, Area2, and Area3, the correctly detected features were 68.98, 78.78, and 69.80 percent, respectively.