Research On Enhancement Algorithm Of Low Illumination Environment Image

As know,image processing is widely used in the monitoring,medical,industrial,and security,among others.But,to improve the accuracy and to apply the effectiveness,the color distortion,low contrast,and detail loss of images acquired under low light need to been enhanced at beginning.Therefore,the algorithm,enhancement,correction,evaluation,and system are studied in this thesis,and the main works and achievements as follow:(1)The research objective,and advancements of enhance processing algorithms are firstly introduced and analyzed for the low illumination image,the thesis topics are proposed with the technical difficulties and problems.(2)A low illumination image enhancement algorithm based on histogram equalization and the activation functions is proposed using the HSV color space.The logarithmic scaling function and hyperbolic tangent function are employed to improve the local contrast of the compressed backlight brightness component,and the bilateral filtering is adopted to eliminate the noise from the backlight brightness component.The histogram equalization,Gompertz distribution,and error function are introduced to enhance the general brightness of low brightness components.Finally,the logarithmic image model is used to fuse the high and low brightness components,and the results are output.Compared to the other algorithms,the algorithm is more subjective and objective in evaluation.(3)A physical model and nonlinear function-based image enhancement method is proposed.The image is converted from RGB to the color space of HSI to extract the reflection component of the I component by using the multi-angle Gabor function.The luminance enhancement model is built on Weber-Fechner law and a logarithmic scaling function to enhance reflection component,and an image fusion strategy is adopted to fuse the reflection components enhanced by different parameters.The test results demonstrate that the technique can enhance the general illumination and contrast within the picture and is good at low-light image enhancement algorithm.(4)A system for image selection,algorithm enhancement,and data evaluation and analysis is designed and developed in this thesis.The feasibility of the proposed algorithm is verified by selecting the target images,enhancing the images using the proposed algorithm,and analyzing enhancement results,and the effectiveness of the system is further confirmed by the developed system.