A novel spectral index to automatically extract road networks from WorldView-2 satellite imagery

• Main Authors: Shahi, Kaveh, Mohd Shafri, Helmi Zulhaidi, Taherzadeh, Ebrahim, Mansor, Shattri, Muniandy, Ratnasamy
• Format: Article
• Language: English
• Published: Elsevier 2015
• Online Access: http://psasir.upm.edu.my/id/eprint/43542/1/A%20novel%20spectral%20index%20to%20automatically%20extract%20road%20networks%20from%20WorldView.pdf; http://psasir.upm.edu.my/id/eprint/43542/
• Institution: Universiti Putra Malaysia

This research develops a spectral index to automatically extract asphalt road networks named road extraction index (REI). This index uses WorldView-2 (WV-2) imagery, which has high spatial resolution and is multispectral. To determine the best bands for WV-2, field spectral data using a field spectroradiometer were collected. These data were then analyzed statistically. The bands were selected through the methodology of stepwise discriminant analysis. The appropriate WV-2 bands were distinguished from one another as per significant wavelengths. The proposed index is based on this classification. By applying REI to WV-2 imagery, we can extract asphalt roads accurately. Results demonstrate that REI is automated, transferable, and efficient in asphalt road extraction from high-resolution satellite imagery.