Order based feature description for high-resolution aerial image classification

We propose to use two intensity order-based descriptors for classification of high-resolution aerial images. By analyzing the characteristics of aerial images, it points out that the imagery does not have an absolute reference frame, and can be seen as a fusion of natural image and texture image. For these characteristics, it uses two intensity order based approaches to extract low-level features of aerial images. The two descriptors are inherently rotation invariant and encode complementary information about the image. It then uses a bag of visual words model to build a holistic representation of images. Experiments results on publicly available aerial scene imagery dataset show that linear combination of order based features encodes complementary information and is significantly better than SIFT. In addition, we make a consistent comparative analysis of different classification frameworks and several feature coding methods based on aerial image dataset.