Object identification and visualization of urban road materials using narrowband near infrared imaging indexes

Classification performance analysis was done on selected narrowband near-infrared spectrum to identify urban road materials. There were five urban road materials to be identified; aggregate, asphalt, carbon organic, clay and concrete. The imaging indexes were selected that proposed a set of narrowband 760nm, 800nm, 850nm, 900nm, 950nm, 980nm with 720nm wide-band spectrum as normalization spectrum. Multilayer perceptron classification method was used. The result was presented where a significant classification performance shown on those five urban road materials with 76.6032 % using 10 fold cross validation on data data samples. A separate test data samples used for further predictions and visualization. Keyword: Imaging indexes, near-infrared, feature classifier, prediction and visualization, multilayer perceptron