Abstrat:By using the cluster analysis method according to the spectral luminous efficiency functions, the linear dimensionality reduction in multispectral dataset was realized, and the independent components of the dimensionality reduction data were extracted by using the fast independent component analysis method.Then, the spectral space was reconstructed according to the independent components.Finally, contrast this method and the principal component analysis method from the mean square error and color difference.The experimental results showed that the average value of mean-square error of proposed method reduced by 3.64% and the average value of color difference increases by 24.08% in comparison with PCA.The multispectral space reconstructed by using hybrid algorithm can efficiently represent the original spectral space.%根据光谱光效率函数,利用聚类分析法对多光谱数据集进行线性降维,进而利用快速独立成分分析法对初次降维数据提取独立成分,然后根据独立成分进行光谱空间重建,最后从均方误差以及色度空间两方面对此方法与主成分分析法进行对比。实验结果表明,该方法的均方差平均值较PCA降低了3.64%,平均色差较PCA降低了24.08%。可见,利用混合算法重建的多光谱表示空间能够更高效地表示原始光谱空间。
展开▼