| High dynamic range(HDR)image reconstruction is an effective image processing technology.Compared with ordinary digital images,HDR images have a wider dynamic ra nge.Therefore,HDR images are widely used in the fields of industrial image processing,machine vision,3D entertainment,aerial image,medical imaging and so on.Because there are always moving objects in the image during the actual synthesis process,the image fusion algorithm that is mainly based on the fusion of the dynamic scene is used to obtain the high dynamic range image without ghosting.The image alignment and fusion function are the research focus about the fusion of dynamic target scene images.This thesis mainly studies the two subjects mentioned above and makes achievements in the process of fusion as follows:Firstly,in order to avoid the formation of ghosting,the method of image alignment is improved based on the two-way similarity of the images.In the process of aligning images,the improved algorithm takes into account the completeness of the image and the consistency between the input image and the reference image,thus ensures the quality of the image after alignment and gets ready for image fusion.Secondly,the function of the fusion function in the fused image is analyzed in the thesis.In order to improve the weight of the reference image,the proportion of reference image is added in the process of image fusion.According to the features of the abnormal region,the cosine function is employed instead of the piecewise linear function.The proportion of overexposure and underexposure in images is reduced and the effect of high dynamic range image is enhanced.Finally,the algorithm is simulated to fuse the sequence images,and the results are compared with those of the Sen P algorithm and Dou Lin algorithm.The simulation results show that the synthesized algorithm is superior to the above two algorithms. |
