Object classification of remote sensing images based on BOV

In order to classify the remote sensing land-use objects efficiently with the very high resolution (VHR) remote sensing images, this letter proposes a remote sensing object classification method based on bag of visual words (BOV) model. This method combines the scale invariant feature transform (SIFT) feature and the texture feature as remote sensing words. Then, the remote sensing words are used for generating word frequency histograms. The histogram is the bridge of the remote sensing words and the classifier. At last, in the classifier design section, the histogram intersection kernel (HIK) is adopted in the SVM. We use the proposed classification method to classify the UC Merced dataset and self-made dataset. Experimental results show that the proposed remote sensing object classification method yields better classification than the existing methods in terms of the classification accuracy.