论文部分内容阅读
针对目前多聚焦图像融合方法处理含噪图像缺乏有效性而导致融合效果较差的问题,提出一种引导滤波结合脉冲耦合神经网络(PCNN)的非下采样Shearlet变换(NSST)域内多聚焦图像融合方法。首先,分别对待融合多聚焦图像进行NSST获取其相应高频子带和低频子带系数;对高频子带系数,通过引导滤波结合改进简化PCNN模型设置融合规则;提取相位一致性、清晰度和亮度等底层视觉特性,指导低频子带系数融合权重;最后反NSST获取最终融合结果。实验结果表明,本文方法能够在噪声干扰情况下有效完成多聚焦融合,并且边缘和纹理信息保持较好,当20标准差噪声时互信息提升了近0.15具有有效性。
Aiming at the problem of the lack of effectiveness of the multi-focus image fusion method to deal with the lack of effectiveness of the noisy image, a new fusion algorithm based on guided filtering combined with pulse-coupled neural network (PCNN) is proposed for multi-focus image fusion in nonsubsampled Shearlet transform (NSST) method. Firstly, the fusion multifocus images are treated respectively by NSST to obtain the corresponding high frequency subband and low frequency subband coefficients; for the high frequency subband coefficients, the fusion rules are set through the guided filtering and the improved simplified PCNN model; the phase consistency, sharpness and Brightness and other underlying visual characteristics, guide the low frequency subband coefficient fusion weight; the final anti-NSST to obtain the final fusion results. The experimental results show that the proposed method can effectively perform multi-focus fusion under the condition of noisy interference, and the edge and texture information remain well. When the standard deviation noise increases by nearly 0.15, the mutual information is effective.