Land-Use Scene Classification in High-Resolution Remote Sensing Images Using Improved Correlatons

Existing methods that incorporate spatial information into a traditional Bag-of-Visual-Words (BoVW) model consider the spatial arrangement of an image but ignore pixel homogeneity in land-use remote sensing images. In this letter, we present an improved correlaton model to jointly integrate appearance, spatial correlation, and pixel homogeneity using multiscale segmentation. The effectiveness of the proposed method was tested on a ground truth image data set of 21 land-use classes manually extracted from high-resolution remote sensing images. The experimental results demonstrate that our improved correlaton model can promote classification and outperforms existing methods such as the traditional BoVW model, spatial pyramid matching model, and the traditional correlaton model.