Pan-sharpening as an effective method to improve classification accuracy of roofing materials
Abstract
In urban environment the presence of hazardous roofing material is a serious threat for both the inhabitants and the environment. Since they provide danger from several aspects: they can be flammable but also can cause cancer like asbestos roof types. A possible way for the mapping of roofing materials is the remote sensing since it can provide high resolution datasets from a secure distance. However, the identification precision of these materials are not always reliable and this research area still needs more improvements. Our aim was to provide an analysis on the classification accuracy of roofing materials using a WorldView-2 satellite imagery covering a selected area of Debrecen, the second biggest city of Hungary. We investigated the Discriminant Function Analysis (DFA) and Random Forest classifications with different class numbers while we also applied a pan-sharpening on the imagery to reveal its possible role in the classification accuracy. Our results showed that the both classifiers effectively identified the roofing materials by the involved classes having the overall accuracy values higher than 85%. RF classifier provided the best results in case of three and six classes, while DFA also showed effective classification in case of three classes. Linear DFA was found to be the worse in case of accuracy however this value generally was around 85%. It was possible to identify the harmful asbestos material using all the classifiers, therefore, we suggest that this roof mapping method can be used by the local management authorities and cadastral plans. The research was supported by the TNN 123457 project. Abriha, D. was supported by the ÚNKP-18-2-I-DE-81 New National Excellence Program of the Ministry of Human Capacities. References: Abriha, D., Kovács, Z., Ninsawat, S., Bertalan, L., Balázs, B., Szabó, Sz. (2018). Identification of roofing materials with Discriminant Function Analysis and Random Forest classifiers on pan-sharpened WorldView-2 imagery - a comparison. Hungarian Geographical Bulletin 67(4), 375-392. Balázs, B., Bíró, T., Dyke, G., Singh, S.K., Szabó, Sz. (2018). Extracting water-related features using reflectance data and principal component analysis of Landsat images. Hydrological Sciences Journal 63(2), 269-284. Enyedi, P., Pap, M., Kovács, Z., Takács-Szilágyi, L., Szabó, Sz. (2018). Efficiency of local minima and GLM techniques in sinkhole extraction from a LiDAR-based terrain model. International Journal of Digital Earth. Szabó, S., Burai, P., Kovács, Z., Szabó, G., Kerényi, A., Fazekas, I., Paládi, M., Buday, T., Szabó, G. (2014). Testing of algorithms for the identification of asbestos roofing based on hyperspectral data. Environmental Engineering and Management Journal 143, 2875-2880.
- Publication:
-
EGU General Assembly Conference Abstracts
- Pub Date:
- April 2019
- Bibcode:
- 2019EGUGA..2114594S