常用的多聚焦融合算法不能很好的将多个聚焦不同图像的清晰部分融合成一幅清晰的图像,提出了一种有效的多聚焦图像融合算法;首先对多聚焦原始图像进行NSCT变换,得到不同尺度与方向下的子带系数;其次利用光学系统成像原理及点扩散函数在光学成像过程中的作用,对低频系数判定多聚焦图像中的聚焦点的近似位置;然后将高频系数作为PCNN的输入,选取区域对比度为PCNN的链接强度,利用PCNN的全局耦合特性和脉冲同步特性选择高频系数;最后经NSCT逆变换得到融合结果;实验结果表明,该方法得到的融合图像在边缘等细节上比常用方法具有更好的视觉效果,在熵、平均梯度、标准差等客观指标上优于小波融合方法和NSCT等传统融合方法。
An improved multi--focus image fusion method based on NSCT and self--adaptive PCNN is proposed. Firstly, we perform the NSCT transform to mu~ti--focus image to ahtain Suhband Coefficients in different subbands; Secondly, to abtain the Focus Point of multi --focus image in low frequency coefficient aecording to the principle of optics and PSF; Then using the high frequency coefficient as the input of PCNN, choosing the contrast among adjacent regions as the linking strength of PCNN, choosing the high frequency coefficient by making use of coupling properties and synchronous pulse properties; Finally , it can obtain a fused image by taking inverse NSCT. The results of simulation and quantifying evaluation show that this algorirhm can preserve more useful information from the original multi--focus image ef- fectively, and enhance the quality of the fused image.