Research Of Multi-focus Image Fusion Based On Non-subsampled Contourlet Transform

Image fusion technology can break through the limitations of the single sensor to obtain image information,extract information from multiple sensors for image fusion,so as to improve the clarity of scene or target in the image.This technology facilitates human visual observation and machine learning,and contributes to the next work and research.Multi-focus image fusion is not only widely used in civil,but also plays an important role in military,agriculture and medicine.Firstly,in view of the large amount of source image information loss in the traditional Non-subsampled Contourlet Transform(NSCT)fusion algorithm,this paper proposes a multi-focus image fusion based on the focus region detection in NSCT transform domain.Compared with the traditional fusion method,this method makes full use of the correlation between the direction sub-bands of each level.On the whole,this method retains the original information of the image to the greatest extent,and improves the mutual information and fusion effect of the image.Secondly,in order to improve the reservation of the original information which is meaningful to the fusion image and reduce the information loss in the fusion method based on transform domain,this paper proposes an image fusion algorithm based on NSCT focus area detection.This method combines transform domain algorithm and spatial domain algorithm,and makes full use of the advantages of transform domain and spatial domain.It not only improves the spatial structure characteristics of the image,but also fully retains the details of the source image.Finally,aiming at the case that the focus degree of the smooth region may be misjudged in spatial domain algorithm,this paper proposes the secondary image fusion algorithm based on NSCT,which uses the Recursive edge-preserving filter(RF)and the Generalized Random Walk(GRW)algorithm in spatial domain image fusion processing.It could improve the accuracy of the focus area detection.After that,NSCT is used to decompose the two fusion images,and the second fusion is carried out.This method makes the image have a more accurate judgment for the focus area,retains the original information of the image to the greatest extent,and enhances the spatial continuity of the fusion image.Experimental results show that this method can effectively improve the quality of fusion image and improve the final fusion image visual effect.