Research Of Wavelet Image Fusion Approach Based On Edge Significance

With the development of multi-sensor technology, image data has increased dramatically. How to utilize image data of different sensors to get more reliable images is becoming a hot research field of image processing. This is precisely the contents of image fusion.The image edge is important for the object recognition in image processing,And as a important domain and procedure in image prcessing,image fusion should keep the edge information of source images as much as possible.There are many image fusion approaches. Wavelet-based image fusion approach is currently an important research direction. Wavelet transformation decomposes an image into low-frequency section containing the image contours and high-frequency section containing the image details and image edges. low-frequency information also plays a crucial role in fusion results.Wavelet image fusion is significantly affected by wavelet parameters. Wavelet parameters should be different when fusing different types of images. In this paper, a series of experiments are performed to determine the selection of wavelet parameters. With the aim of keeping edge information as much as possible in the fusion images, a series of researches have been done in this paper. Firstly, characteristics of image edge are analysed. Secondly, model of edge significance is established. Finally, a wavelet image fusion approach based on edge significance is proposed.As is shown in the experiments, this approach can get more satisfied results in keeping the edge information of source images.