三种像素级数据融合算法分析与性能比较

Comparisons and Performance Analysis on Three Pixel level Data Fusion Algorithms

摘要:

讨论了加权均值法、彩色空间变换法(HSI)和主成分分析法(PCA)三种像素级数据融合算法及实现, 然后对全色图像和多光谱图像做了数据融合仿真实验, 最后从信息熵、清晰度和光谱保持特性三个方面进行了性能综合评价。实验结果表明, 基于HSI变换的融合图像整体清晰, 色度协调, 保留了较多的空间信息, 细节特征明显, 质量最好, 能够满足实际应用平台的技术需求。

Abstract:

Discussed were three kinds of pixel level data fusion algorithms and their implementations, including weighted mean method, hue saturation intensity method (HSI) and principal component analysis (PCA). Then data fusion simulation experiments were carried out on full color images and multispectral images. Finally, comprehensive assessments were performed on the performance from the three aspects of information entropy, definition and spectrum feature. The experimental results show that the blending images based on HIS are clear, coordinative in color, obvious in detail features, and reserve more spatial information.