Hyperspectral image classification with limited number of labeled pixels is a challenging task. In this paper, we propose a bilayer graph-based learning framework to address this problem. For graph-based classification, how to establish the n

本文提出的超像素渐进合并算法应用了图论的基本框架：一幅图像被映 射成一个加权的无向图，其中超像素被当作图中的顶点，而相邻超像素之间通过边 相连，由超像素间的相似度计算出相应边的权值，通过排序每次合并权值最小的两 个超像素。-Shot into a weighted undirected graph, including super pixels as the vertices in the graph, and the adjacent pixels by boundary betw