Detection of Damaged Areas Based on Visual Bag-of-Words Model from Aerial Remote Sensing Images

An approach for damaged rooftops areas detection is proposed based on visual bag-of-words model. First, the building rooftop is segmented into different superpixel areas using simple linear iterative clustering(SLIC) method, then features of color and histograms of oriented gradients are extracted from each superpixel area and the visual bag-of-words (BoW) model is employed to build the semantic feature vectors of damaged and non-damaged area. Finally, damaged and non-damaged parts of rooftop superpixel areas are discriminated using SVM. Experimental results show that the proposed method can be feasible and effective for detection of damaged rooftop areas, which is an important significance for improving the accuracy of overall building damaged detection.