摘要
结合空间信息和投票机制,提出了将ISODATA与K-means投票融合的空间预处理方法用于端元提取,并结合3种经典的端元提取算法N-Finder、VCA和OSP进行验证。通过模拟数据与矿区真实数据的对比实验,验证了加入空间信息的端元提取算法能够更有效抑制噪声,从而使提取的端元光谱更加准确。
In this paper, considering the majority voting mechanism and spatial information, we used a spatial preprocessing method integrating ISODATA with K-means voting to extracte endmember, and used three classical endelement extraction algorithms, such as N-Finder, VCA and OSP to verify the results respectively. Then, we conducted the experiments on one simulation data set and one real data set. The results show that the proposed method which considering the spatial information can restrain the noise effectively and achieve better accuracy for endmember extraction.
出处
《地理空间信息》
2020年第2期34-37,I0006,共4页
Geospatial Information
基金
2017年四川省科技计划项目(2017SZ0081)。
关键词
高光谱影像
端元提取
空间预处理
hyperspectral image
endmember extraction
spatial preprocessing
