论文部分内容阅读
针对Isomap-NFINDR端元提取算法复杂度高、占用内存多、效率低的缺点,提出一种基于标志点选择Isomap的快速端元提取算法。该方法采用最大最小距离法来选取初始的K个聚类中心点,并采用光谱夹角距离SAD代替欧式距离来进行聚类分割;根据图像的空间特性,从去除聚类的边界点后剩余点间隔抽取距离聚类中心距离最小的N个点作为标志点。真实高光谱图像实验结果表明,提出的算法精度接近原始的基于Isomap-NFINDR算法,而效率提高了将近60倍。
Aiming at the disadvantage of Isomap-NFINDR algorithm, such as high complexity, high memory consumption and low efficiency, this paper proposes a fast endmember extraction algorithm based on marker selection Isomap. This method uses the maximum and minimum distance method to select the initial K cluster centers, and uses spectral distance SAD instead of the Euclidean distance to carry out cluster segmentation. According to the spatial characteristics of the image, after removing the cluster point, At intervals, N points with the smallest distance from the cluster center are taken as mark points. Experimental results on real hyperspectral images show that the proposed algorithm is close to the original one based on Isomap-NFINDR algorithm, and the efficiency is improved nearly 60 times.