小波融合图像效果的因子分析评价方法

因子分析是一种有效的评价融合图像质量的方法。文中采用了小波sym5进行了多尺度图像融合的研究，并且通过主成分分析方法提取了细节信息因子和光谱信息流失因子等参数，作为图像的客观质量评价标准。实验证明，对于小波变换多尺度融合后的图像，该方法能够客观有效地评价其反映细节信息的参数（如信息熵和清晰度）以及反映光谱信息的参数（如扭曲程度和偏差指数）的变化趋势。理论分析结果，因子综合得分和观众评价得分结果具有良好的一致性。

Quality Evaluation of Wavelet Fused Image by Using Factor Analysis Method

Factor analysis is a powerful way in the field of evaluation of fused image quality. In this paper, wavelet sym5 is used to decompose and construct image at different level. Factors of detail information and spectrum information lacking extracted by PCA(principal component analysis) are selected to objectively evaluate the fused image. Experimental results show that this approach can effectively analyze the variation trend of detail information(such as entropy and definition) and spectrum information(such as Distortion coefficient and Deviation coefficient) of the fused images. Variation of factor integrated score and audience evaluation effect is the same. Therefore theoretic analysis results and visual effect of fused image are in accordance with effect analyzed by using factor analysis approach.