基于提升小波变换的图像融合改进算法的研究

Abstract

提出了一种基于提升小波变换的图像融合改进算法。针对提升小波分解后的低频和高频分量各自的特点，选用不同的规则进行融合，即低频系数采用选择法和加权平均相结合的策略，高频系数时，把小波系数的方差与绝对值综合起来考虑决定融合小波系数。实验结果表明，当采用平均梯度、信息熵、标准差、均方根误差和峰值信噪比作为客观评价准则，该算法的融合图像比拉普拉斯金字塔融合图像和传统的小波变换的融合图像具有更好的融合效果，较好地提高了图像融合精确度。 An improved algorithm is Proposed for image fusion based on lifting wavelet transform in the paper. According to the characteristics of lifting wavelet decomposition of the low and high frequency components of the respective, different fusion rules are adopted, namely low frequency coefficient selection method and the weighted average method, choosing the high frequency coefficient, the variance of wavelet coefficients and the absolute value of the wavelet coefficients are considered synthetically decision fusion. The experimental results show that, when we take the average gradient, the information entropy, the standard deviation, root mean square error and peak signal to noise ratio as the objective evaluation criteria of image fusion, image fusion algorithm has better fusion effect than Laplasse Pyramid fusion image and the traditional wavelet transform; it improves the accuracy of image fusion effectly.