Reaserch On Gradient Histogram Equalization Based Image Enhancement Algorithms

Image enhancement is an important direction of image processing.Its goal is to highlight important areas in the image and increase the overall and local contrast of the image.Histogram equalization is an effective method for image enhancement due to its good real-time performance and obvious enhancement effect.The image gradient can measure the rate of change of the grayscale of the image,and its main reaction is the dynamic information of the grayscale and the texture details of the image.If enhancing the image gradient field,then it can make the details of the image clearer,therefore,the enhanced image gradient field can reduce the impact of lack of texture details due to factors such as lighting conditions to a certain extent.In short,it is of great significance to study how to effectively equalize or transform the gradient field to enhance the essential texture details of the image.This thesis focuses on image enhancement algorithms based on gradient histogram equalization.The main innovations achieved are as follows:1.Aiming at the problem of gradient field over-enhancement and gradient level phagocytosis of current algorithms,this thesis applies the histogram equalization method to the gradient field,and proposes a gradient field equalization algorithm based on edge segmentation and probability correction.The algorithm regards the gradient distribution of the original image as a mixed Gaussian distribution,and uses the adaptive probability statistical method to divide the gradient field of the original image into edge and non-edge parts.Then each part is independently equalized to improve the over-enhancement effect of current algorithms.In addition,before the equalization of the edge and non-edge parts,the single threshold method is used to modify the probability of the two parts to avoid the gradient level phagocytosis effect of histogram equalization.The simulation results show that the proposed algorithm increases the average gradient of the image by up to three times,compared with the comparison algorithms,the enhanced image is not distorted.2.Aiming at the problem of high time complexity and insufficient image quality of image reconstruction by current algorithms,on the basis of the algorithm proposed in 1,this thesis further proposes a gradient field equalization algorithm based on matrix replacement method reconstruction.The algorithm uses the algorithm proposed in 1 to obtain the target gradient field.Then,when reconstructing the enhanced image based on the target gradient field,the matrix substitution method is used to find out the Laplace operator and the divergence operator of the image matrix,and the solution of the Poisson equation is converted into a matrix operation solution to improve the reconstruction time complexity and image quality.The simulation results show that: compared with current reconstruction method,the proposed matrix replacement method reduces the reconstruction time complexity and retains the details of the enhanced image;compared with the algorithm proposed in 1,the algorithm time complexity is improved.