模糊统计耦合子直方图加权均衡化的图像增强

Image enhancement based on fuzzy statistical coupling sub-histogram weighted equalization

基于直方图均衡化的图像增强算法易出现过渡增强与棋盘效应,难以保持图像亮度,导致失真等不足,为此提出一种模糊统计耦合子直方图均衡化的抗失真图像对比度增强算法。引入模糊集理论,将图像转变成模糊矩阵;通过隶属度函数与图像灰度水平呈现的概率,嵌入加权函数,构造加权模糊直方图计算模型;利用初始图像的中值,将模糊直方图分割为两个子直方图,定义累积密度函数,构造其变换模型;建立逆变换函数,对其完成解模糊化,输出增强图像。实验数据表明,与当前基于直方图均衡化的图像增强算法相比,该算法显著消除了过度增强与噪声放大,具有更佳的增强视觉质量与抗失真性能,其平均信息内容AIC(average information contents)值最大,自然图像质量评估值NIQE(natural image quality evaluator)最小。

To solve these defects such as difficult to maintain image brightness,prone to transition enhancement and checkerboard effect resulting in distortion in current image enhancement algorithm based on histogram equalization,the anti-distortion image contrast enhancement algorithm based on fuzzy statistical coupling sub-histogram equalization was proposed.The image was transformed into fuzzy matrix using the membership function and the pixel intensity.The fuzzy histogram calculation model was constructed by introducing the fuzzy set theory.The fuzzy histogram was divided into two sub-histograms,and transformation model was designed by defining the cumulative density function.The inverse transformation function was established to defuzzify for outputting the enhancement image.The experimental results show that this algorithm significantly eliminates the excessive enhancement and noise amplification with better visual quality and anti-distortion performance for the maximum value of average information content and minimum value of natural image quality assessment.