Title :
Urban Land-Use Multi-Scale Textural Analysis
Author :
Pacifici, F. ; Chini, M. ; Emery, W.J.
Author_Institution :
Earth Obs. Lab., Tor Vergata Univ., Rome
Abstract :
Urban areas are composed of numerous materials arranged by humans in complex ways. A simple building may appear as a complex structure with many architectural details surrounded by gardens, trees, buildings, roads, social and technical infrastructure and many temporary objects, such as cars, buses or daily markets. In this paper, we analyze the effectiveness of 8 textural features (resulting from the Grey Level Co-occurrence Matrix) derived from a 50 cm WorldVieW-1 image of Washington D.C. (U.S.A.). The information extracted from the panchromatic and textural features are fused and processed by a Multi-Layer Perceptron (MPL) neural network producing a land-use map with accuracy above 0.90 in term of K-coefficient.
Keywords :
geophysical signal processing; image texture; multilayer perceptrons; neural nets; remote sensing; USA; Washington D.C; WorldVieW-1 image; architectural details; buildings; complex structure; gardens; grey level co-occurrence matrix; land-use map; multilayer perceptron neural network; multiscale textural analysis; panchromatic images; roads; social infrastructure; technical infrastructure; trees; urban land use; Automotive materials; Buildings; Data mining; Humans; Image analysis; Image texture analysis; Multilayer perceptrons; Neural networks; Roads; Urban areas; Multi-scale texture analysis; WorldView-1; urban land-use;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 2008. IGARSS 2008. IEEE International
Conference_Location :
Boston, MA
Print_ISBN :
978-1-4244-2807-6
Electronic_ISBN :
978-1-4244-2808-3
DOI :
10.1109/IGARSS.2008.4780098
