具有视觉特性的局部直方图均衡增强算法

A Local Histogram Equalization Algorithm Utilizing Visual Characteristics

直方图均衡能有效地对数字图像进行增强处理,其主要缺点是处理后的数字图像大量损失灰度层次,且增强力度不够.提出了在子块部分重叠直方图均衡的基础上,引入保留灰度级直方图均衡处理的方法,使得处理后的图像对比度得到较大的增强,又不损失灰度级,且相邻像素间灰度过渡平缓,图像柔和、真实.针对处理后方块效应严重的图像,依据人眼的视觉特性和数字图像数据的特点,提出了自适应行列分比例方法来有效消除块效应.模拟结果显示,效果显著.针对子块过增强带来麻点噪声的缺点,提出了根据子块方差来判断对比度拉伸强度的方法,一定程度上削弱了麻点噪声对视觉效果的影响.

Digital images can be effectively enhanced by using histogram equalization. However the main limitations of the algorithm are the significant loss of gray levels and the lack of enough contrast enhancement ability in the processed image. In this paper, an algorithm of graylevelpreserving histogram equalization is introduced, which is performed on partially overlapped blocks. The simulation result shows that this method not only enhances the contrast, but also preserves the gray information. Furthermore, no sudden gray level changes occur between the adjacent pixels so that the processed image looks smooth and realistic. A novel method is also presented according to the image data and the human visual characteristics, which is used to effectively reduce the block effects caused by the block based histogram equalization. The simulation result shows that the improvement is obvious. To eliminate noise effects caused by overenhancement, another measure is taken to make the strength of contrast enhancement vary with the variance of subblocks. Experiments show that noise is reduced to some degree.