Advanced search
Start date
Betweenand

Application of Support Vector Machine in airborne laser scanning data and aerial images for the impervious surfaces classification

Grant number: 16/05240-8
Support type:Scholarships in Brazil - Scientific Initiation
Effective date (Start): May 01, 2016
Effective date (End): April 30, 2017
Field of knowledge:Physical Sciences and Mathematics - Geosciences - Geodesy
Principal researcher:Tatiana Sussel Gonçalves Mendes
Grantee:Andressa Nalú Hernandes
Home Institution: Instituto de Ciência e Tecnologia (ICT). Universidade Estadual Paulista (UNESP). Campus de São José dos Campos. São José dos Campos , SP, Brazil

Abstract

Researches on the mapping of impervious surfaces from remotely sensed data has attracted interest since the 1970s. In the urban context, the study of impervious surfaces is highly relevant and can contribute to urban and environmental planning and resource management. The available of high resolution images has motivated and increased the interest in studies that focus on urban environment, but the high detail provided by these images present challenges for image processing techniques in order to classify urban scenes, such as: confusion in spectral response between different types of land cover; high variability of pixels within the same land cover class; and shadows caused by aboveground objects. One way to overcome these challenges is incorporate additional information in the classification process, for example, Airborne Laser Scanning (ALS) data. In this sense, this research proposes the application of Support Vector Machine (SVM) for to classify urban impervious surface from the integration of ALS data and high resolution aerial images.