Weighted Generalized Nearest Neighbor for Hyperspectral Image Classification

In this paper, we develop an effective classification framework to classify a hyperspectral image (HSI), which consists of two fundamental components: weighted generalized nearest neighbor (WGNN) and label refinement. First, we propose a novel WGNN method that extends the traditional NN method by introducing the domain knowledge of the HSI classification problem. The proposed WGNN method effectively models the spatial consistency among the neighboring pixels by using a point-to-set distance and a local weight assignment. In addition, we develop a novel label refinement method to enhance label consistency in the classification process, which is able to further improve the performance of the WGNN method. Finally, we evaluate the proposed methods by comparing them with other algorithms on several HSI classification data sets. Both qualitative and quantitative results demonstrate that the proposed methods perform favorably in comparison to the other algorithms.