Abstract: Roads are the traffic lifeline for emergency rescue and disaster relief. After huge earthquakes, it is very important to rapidly extract road damages from remote sensing images for disaster emergency rescue. However, because damages caused by earthquakes is different at different places, it is very complicated to extract in a common way. Usually, spectral features are used to discriminate damaged and undamaged pixels or objects in images. Because thresholds have to be chosen by manual operation dependent on different situations, methods based on spectral features are low in accuracy and efficiency. This paper proposes an approach to rapidly extract road damages by earthquake based on sidelines. The approach extracts double sidelines of undamaged road from images acquired after earthquakes, and then compare them with vector road lines before earthquakes. The damaged segments of roads can be extracted through comparison. The performance of the method is evaluated by an experiment with QuickBird images in the disaster area of the 2008 Wenchuan earthquake.

Key words: Sidelines, Road Damage, Extraction, Remote Sensing