Pothole Detection Based on Superpixel Features of Unmanned Aerial Vehicle Images

Potholes are a significant expressway engineering disease. Automatic inspection based on unmanned aerial vehicle (UAV) images can identify potholes more efficiently than manual inspection. However, few pothole recognition methods are based on UAV images. A pothole recognition algorithm based on the superpixel features of UAV images is proposed to efficiently detect potholes. The shape, color, and texture of the potholes are utilized in the different steps of this method. First, the image is transformed into the hue–saturation–value (HSV) color space to detect the background according to its colors. Second, the pixels are partitioned into superpixels based on their similarity, and the gray co-occurrence matrix is used to calculate the texture features of the superpixels to obtain the feature vectors. Finally, the superpixels in the potholes, are detected based on the clustering of the feature vectors. The proposed method is compared with a state-of-the-art method to verify its performance. The relative error of the proposed method is at most 4% for most images. The experimental results demonstrate that the proposed algorithm can detect potholes in UAV images more accurately. The proposed method is practical and reproducible for automatic inspection of expressway diseases using UAV images.