K-Nearest Neighbor combined with guided filter for hyperspectral image classification

Explosive growth of applications in hyperspectral image has brought challenge of how to efficiently classify the objects by their spectral feature. Under this circumstance, to improve the classification accuracy, lots of spectral-spatial approaches are adopted, instead of traditional pixel-wise classification. In this paper, we combine k-nearest neighbor with guided filter to mine spatial information effectively or and optimize the classification accuracy. To verify the feasibility of the two proposed methods, we evaluate performance over two benchmark datasets. Comparative experiments suggest that the proposed approaches show better accuracy.