Abstract: To extract the salient features of the image more effectively and improve the performance of multi-focus image fusion，a new method of image fusion is proposed in this paper.For the defect of image feature can’t be extracted by filter generated through the use of Gaussian kernel basis.In this paper，exploiting the advantages of approximate orthogonal of wavelet kernel and signal local analysis for support value transform，Pulse-Coupled Neural Network is applied for the low frequency information，while the maximum absolute value selection fusion rule for the high frequency information.The experiment data and theory analysis show that this method effectively improves the salient features of the image.The information entropy evaluation results etc，are raised and visual effects is improved than the method using support value transform with Gaussian kernel.

Key words: wavelet kernel, support value transform, Pulse-Coupled Neural Network（PCNN）, the maximum absolute value selection fusion rule

摘要： 为更有效地提取图像的显著特征，提高多聚焦图像融合的性能，针对高斯核不完备基的缺点，其生成的滤波器不能有效提取图像显著特征，利用小波核近似正交和信号局部分析的优点，构造支持度变换，经过支持度分解后的低频信息使用PCNN的融合规则，高频信息使用绝对值最大选取的规则进行图像融合，实验数据和理论分析表明：该方法有效地改进了图像的显著特征，与高斯核构造的支持度变换图像融合方法相比，信息熵等评价指标结果均有提高，并且视觉效果有所改进。

关键词: 小波核, 支持度变换, 脉冲耦合神经网络, 绝对值最大选取融合规则