High Dynamic Range Image Synthesis Based On Image Fusion

Due to the limitation of camera imaging principle,even in the same light source,a lot of different material article also presents different exposure state,including the image,owe exposed the serious influence the image quality,limits the processing of late.High dynamic range image as a real reduction of natural scenes,has important position in image processing,in transport,security,industrial,and other fields have broad application prospects,and as a result of industrial production line,the synthetic image quality requirements at the same time,also have further restrictions on the efficiency of its time.Currently,the method of synthesizing high dynamic images is divided into two types,based on the calibration of the camera response curve and the direct fusion based on the sequence of multi-exposure images.Based on the calibration of camera response curve method is the theory basis of unified,standardized,systematic,fixed computing model,restore the image of a true care from the principle,but the large amount of calculation and recover the camera response curve of time-consuming;Based on multiple exposure image sequence of direct fusion jumps over physical reduction steps,so simple,but is limited by the selection of image evaluation quality standard,method and varied.But these two methods are still too complex and time-consuming to produce a single line of high speed production and product exposure.So this paper,based on the Mertens algorithm,simplifies the calculation of the Laplace pyramid,and optimizes the time when the quality of the image is guaranteed.Is the pyramid itself complex structures,the bartender while denoising directly the use of bilateral filter to replace the pyramid structure,the weights of smoothly,considering the relationship between space and pixel domain,quickly the fusion image and verify the synthesis effect.Through experimental verification and practical application,the simplified pyramid structure and the method of bilateral filtering can meet the requirement of rapid production line.The main features of the paper innovation is:This paper introduces the high dynamic range image synthesis based on camera curve calibration and image fusion based on image.The basic process and theoretical basis of the two methods are briefly introduced,and the synthetic process is demonstrated by the experiment of the system as the background theory of this paper.The Mertens algorithm based on the Laplacian pyramid has been improved.For the highly productive production requirements and simpler image synthesis requirements in industrial applications,the pyramid decomposition and synthesis algorithms have been simplified to ensure the image quality,Reducing the synthesis time.proposed the use of filters to smooth the weight matrix synthesis method,the use of edge-preserving denoising of the bilateral filter,consider the image pixel spatial distance and range relationship,smooth the weight matrix,and retain the edge of the weighted graph,Improve the time efficiency.