Fusion Algorithm. Chunked Image Registration Computing Research

Image fusion is a modern high technology that integrates many sciences, such as sensor technology, image processing, computer engineering and artificial intelligent. The main purpose of image fusion is to enhance reliability of the image through processing redundant data between many images and to enhance clarity of the image through processing supplementary information between many images. The image fusion technology occupies important places in the fields of remote sensing, medicine, natural resource exploration, terrain landform analysis and biology. Image registration is a crucial step in image fusion.As the image size increasing, the numerical approaches based on the pixel matrix may step-up the computation complexity, and it is difficult to perform the large size registration on a single computer in a limited time. If we do process the whole image, it maybe spend a lot of unnecessary time and won't achieve the accuracy, especially we're only interest in one or several objects in the multi-objectives image. One of the most efficient solving methods is blocked registration.The crucial step is merged blocks after finish the registration, so it need to image fusion. We briefly discuss the several performances of the image fusion quality evaluation. Then the choice of the wavelet and the decomposition level and the fusion ruler for the Wavelet transformation based on the image fusion method are discussed. The algorithms include:average rule, area energy rule and area variance rule. Aim at the average rule of the low frequency and the area variance rule of the high frequency. The experimental results show that the fused image produced by the proposed method is better.