Research On Nonlinear Low Light Level Image Enhancement Algorithm

With the continuous development of current information technology,digital images have become an important way for humans to obtain external information.At present,in daily life,the conditions for humans to collect images are often not ideal.In harsh conditions such as rain,snow,or fog.In the external environment,images that cannot clearly express information will be collected,and if the shooting level is low or the shooting machine is poor,the collected images will also be difficult to use.Such images often have characteristics such as low average brightness,blurred details,and large signal-to-noise.They are called low-light images.Low-light images must be enhanced by image processing to meet daily life or scientific research requirements.Retinex theory is an enhancement theory based on the human visual system,which has an ideal effect on improving image quality.It is the focus and hot spot in the current research on low-light image enhancement algorithms.In this paper,a new image enhancement algorithm based on adaptive MSR algorithm and exposure fusion method is proposed for the image with low overall brightness.Firstly,the image is converted into color space to avoid losing color information to the greatest extent,and then the brightness components of the image are segmented according to the requirements.Information entropy is introduced to distinguish the local areas of image with different characteristics.According to the local information entropy of image,a concept of measuring the degree of information richness is defined.Then,the scale parameters of Gauss kernel in each local area of image are calculated respectively to achieve the most ideal enhancement results in the areas with different characteristics of image.To be processed is introduced into the guide filter to retain the structure information of the image.Finally,the exposure fusion method is used to merge multiple processing results,and according to the actual image enhancement experiment and objective image evaluation data analysis,the new image enhancement algorithm based on adaptive MSR algorithm and exposure fusion method is compared with several existing algorithms,and the new image can take into account the advantages of multiple processing methods.In this paper,an image enhancement algorithm based on adaptive logarithm processing model and Retinex theory is proposed for images with low local brightness.Firstly,the algorithm proves that the adaptive PLIP model can meet the needs of image enhancement algorithm in theory,and improves the PLIP model according to the local characteristics of the image.Firstly,the algorithm builds the adaptive logarithm processing model according to the characteristics of different local areas of the image.In this model,low light level image enhancement is enhanced in many aspects,and the result of detail enhancement is processed to maintain the brightness,so as to avoid the over enhancement of the algorithm.Compared with the existing algorithm,the image enhancement algorithm based on the adaptive logarithm processing model and retex theory can effectively solve the problem of information loss in the process of low light level image processing with bimodal characteristics.