Enhancement Of Low Contrast Degraded Image

As we all know,the quality of the image is very important for the use of it.In most cases,the original image acquired by the device could meet the requirements.However,in some cases the acquired image cannot meet the demands,for example,the quality of the image collected in bad weather will be degraded which would cause large amount of information lost as well as make it hard to process and analyze the image.The wildly use of image information has also made this issue more and more noticeable.For example,image information is used in some fields like public transport,military reconnaissance and security In order to solve the problems above,some researches about the image collected in haze and low light conditions has done which aim at defogging and enhancing these images.The main researches are as follows:First of all,the research status and background of image defogging and the enhancement of low light images is introduced.Secondly,several commonly used image algorithms which about image defogging and enhancement of the low light image are introduced respectively.The dark channel prior algorithm and the low-light image enhancement algorithm based on internal noise induction are mainly analyzed and implemented.Moreover,the results were processed and analyzed.In order to overcome the shortcomings of existing algorithms,some corresponding improved algorithms are proposed.For image defogging,an optimized image dehazing algorithm which use guided filter is proposed,it uses the idea of dark channel prior to find the transmission image,and then filter the transmission image by the guided filter to obtain a more refined one.Under the circumstance,the output defogged image would be more smooth and delicate,and the speed of the algorithm has greatly improved by using the color reduction method.As for the enhancement of low-light images,an optimized low-light image enhancement algorithm based on internal noise induction is proposed.The optimized algorithm still use the method of separating the V component in HSV image space then transferring it to the frequency domain and using iteration to enhance it.But the difference is that the parameters of image quality evaluation in the original algorithm are adjusted,furthermore,the conditions to stop the iteration and the iteration step size are refined.The computational complexity of the optimized algorithm is greatly reduced,and because the selection of parameters is more refined,the optimized algorithm could obtain more effective results.Experiments show that the optimized algorithm can be applied to the real-time video processing.