Research And Implementation Of Fast High Dynamic Range Image Fusion

With the development of the times,people began to use more screens,cameras,photos,etc.to observe and record the world,and the demand for high-quality pictures is also growing.However,for real scenes in the real world,its dynamic range can reach dozens of orders of magnitude.At the same time,the current image acquisition equipment is restricted by hardware and can only display a limited dynamic range.It is difficult to take a complete picture of the photos taken to reflect the real scene information.Therefore,high dynamic range image fusion technology came into being and has attracted more and more scholars and experts’ attention.At present,the fusion of high dynamic range images is mainly through collecting multiple photos of different exposures in the same scene,fusing them with multiple exposures,and finally obtaining a high dynamic range image with sufficient information.In this paper,the fusion of high dynamic range image in static scene and dynamic scene is researched and processed.In static scenes,a fusion method based on adaptive weights is used to improve the quality and texture information of the fused pictures.In dynamic scenes,a fusion method based on background modeling is used to remove Ghosts in the process of picture fusion.The main work of the paper is:·This paper systematically introduces high dynamic range image synthesis based on camera curve calibration and direct multi-exposure image fusion.This paper summarizes the various traditional high dynamic range quality evaluation methods that have been proposed,then introduces the current relatively new mainstream quality evaluation methods.·In static scenes,two adaptive weight functions are designed,based on pixel intensity and weight gradient,respectively,which improves the fusion effect of multi-exposure pictures and enriches the information in the sequence of multi-exposure pictures·In dynamic scenes,the background-based modeling method is used to generate a background with almost no moving objects,which better captures the moving objects in the picture,solves the problem of ghosting caused by direct fusion in dynamic scenes,and removes the generated pictures ghost in the picture,improve the quality of the fusion picture·The paper evaluates the algorithm results from both subjective and objective aspects.The experimental results show that the algorithm proposed in this paper can expand the dynamic range of the image and generate high-quality images with good visual effects effectively.