Research On Low Illumination Image Fusion Algorithm Based On Virtual Exposure Technology

Image enhancement technology as part of machine vision plays a critical role in video surveillance,traffic management,location monitoring and other fields.The images taken at night or under uneven lighting conditions have the problems of low contrast,large dark areas,and lack of color,which cannot meet the human visual perception and computer processing requirements.According to the principles of virtual technology and image fusion technology,a low-illumination image fusion algorithm based on virtual exposure technology is proposed in this paper.This algorithm can solve the problems of poor contrast,low definition,and color distortion in traditional algorithms.And a processing website for low-illuminance images is developed,which provides a variety of enhanced algorithms that can quickly process low-illuminance images in real time.Firstly,the original image is converted to a HSV image by projecting onto a HSV color space,the HSV image is separated into a luminance image and chrominance images,and the luminance image is converted to the virtual exposure image set by the virtual exposure technology.Secondly,the virtual exposure image set is denoised and edge processed using the guided filter.The images in set are then divided into blocks,the optimal block of each image is extracted based on local variance,and all optimal sub-blocks are fused using image fusion technology.Finally,the overall brightness of the newly fused image is improved by using gamma correction,resulting in the enhanced image.The experimental results show that the algorithm proposed in this paper increases the mean value of brightness by 127.36%,which effectively improves the clarity of image details and textures,reduces color distortion,and has better subjective and objective enhancement effect.