An image annotation approach using location references to enhance geographic knowledge discovery

► We applied local scale-invariant features and bag-of-keypoints for feature extraction of geographic images. ► We utilized DoG and SIFT methods to extract features for clustering. ► We constructed a visual vocabulary by the AP clustering method, and generated feature vector for each geographic image. ► We used FAM technique to annotate images by using geographic nouns. ► The results of the work can enhance geographic knowledge discovery.