基于权重约束决策的图像增强算法

An Image Enhancement Algorithm Based on Weighted Constraint Decision

为解决当前基于直方图均衡化的图像增强算法易改变初始图像的亮度均值,使其输出图像存在冲蚀效应与棋盘效应的不足,提出了权重约束决策耦合内聚Ostu阈值分割的图像增强算法。首先,建立最优内聚Ostu阈值模型,将输入图像分割为2个子图像,形成目标的低灰度位与背景的高灰度位;再建立低灰度位与高灰度位的直方图均衡化模型;随后,对均衡化子图像赋予融合权重因子,输出增强图像;并引入界限视觉偏差与对比度,建立混沌粒子群算法的目标函数,对权重约束决策的控制参数完成优化。实验结果显示：与全局直方图均衡化、局部直方图均衡化算法相比,所提算法的增强视觉舒适度最好,更好地保持了图像亮度,其输出图像的清晰度值最大。

In order to solve the defects such as wash-out effect and checker-board effect of output images induced by changing the mean brightness of the initial image in image enhancement algorithms based on histogram equalization, the low contrast image enhancement optimization algorithm based on the Ostu threshold segmentation and weighted constraint decision is proposed in this paper. Firstly, the input image is divided into two sub-histograms by defining the optimal Ostu threshold model for forming the low gray level of the target and the high gray level of the background; Then the low gray level and high gray level histogram equalization model is established; the control parameters of the weight constraint decision are optimized by introducing the particle swarm optimization algorithm to determine the optimal constraint for optimizing the enhancement image. Experimental results show that： the proposed algorithm has the best enhanced visual comfort and maintains the better image brightness, and the discrete entropy of the output image is bigger.