有限离散剪切波域结合区域客观评价的图像融合

Image fusion based on area objective assessment in finite discrete shearlet transform domain

为了提升多源图像融合精度,提出了一种有限离散剪切波(FDST)域结合图像区域客观评价的自适应融合方法。该方法利用有限离散剪切波(FDST)对源图像进行多尺度、多方向分解,低频子带图像采用梯度信息相关性因子作为系数权值,高频子带图像应用绝对值与区域标准差一致性选择的融合策略。应用有限离散剪切波逆变换重构图像,采用多组多源图像进行融合试验,并对融合结果进行了客观评价。试验结果表明,本文提出的融合方法在主观和客观评价上均优于其他多尺度分解(MSD)融合方法。

To enhance the accuracy of multi-source image fusion,an adaptive fusion method is proposed.This method is based on the area objective assessment in Finite Discrete Shearlet Transform（FDST）domain.The source images are decomposed to subband images with multi-scale and multi-direction by FDST.The gradient information correlation factor is taken as the coefficient weight for low-frequency subband fusion;while for high-frequency subband,consistency selection of the coefficient absolute value and area standard deviation is adopted as the fusion rule.The fused low and high frequency coefficients are reconstructed to image by Finite Discrete Shearlet Inverse Transform（FDSIT）.Fusion experiment is done with several sets of different modality images,and objective performance assessments of the fusion results are implemented.Results indicate that the proposed method performs better in subjective and objective assessments than other existing Multi-Scale Decomposition（MSD）fusion techniques.