基于成像机理的小波包变换多聚焦图像融合
Multifocus Image Fusion Based on Imaging Mechanism with Wavelet Packet Transform
-
摘要: 由于可见光成像系统的聚焦范围有限,因而在成像过程中,除聚焦良好的物体能生成清晰的图像外,该物体前后一定距离外的所有物体都将呈现不同程度的模糊.为了获得场景内所有物体均清晰的图像,在分析了多聚焦图像成像机理的基础上,提出了一种基于小波包变换的融合方法.它是将成像系统先聚焦在一部分对象上,得到其清晰的图像;然后再将其聚焦在另一部分对象上,得到另一清晰的图像;最后把这两幅实验图像加以融合,从而获得场景内所有物体均清晰的图像.实验结果表明,基于小波包变换的融合方法能够将信号的频带进行多层次划分,对高频成分也能进一步地分解,可有效综合多聚焦图像.Abstract: The multifocus image focus is discussed. Since light imaging system has a limited focus range,all other objects have blurry images to some different degrees besides the focused object,but we often want to acqurire a image on which all objects are clear. In order to get such a image, image fusion method can be applied and two multifocus image can be combined. The property of system function of multifocus imaging system is analyzed according to the circuit and system theory, and the conclusion is that the system function of the optical imaging system is a low pass filter.A multifocus images fusion method based on wavelet packet transform is presented according to the fusion perpose and the property of wavelet packet transform.Two practical multifocus images are fused with the fusion method and a satisfactory clear image is obtained. The experiment results show that the fusion method based on the wavelet packet transform can analyze both the low frequency and high frequency parts of the signal simultaneously at multifocus images.An image on which all objects are clear is provided.