一种基于修正直方图的图像增强算法

Image Enhancing Algorithm Based on the Generalized Histogram

针对雾天降质图像对比度低的问题,提出了一种基于修正直方图的图像增强算法。统计图像的直方图时,每个像素分为两部分,一部分累加到当前像素灰度级,剩余部分按灰度级平均分配,实现直方图的修正,然后根据修正直方图产生灰度映射函数,由于每个像素只有一部分累加到当前灰度级,这样可以避免局部图像的过度增强。实验表明,该算法比经典的直方图均衡化算法、局部直方图均衡化算法有更好的增强效果。

According to the low contrast of fog-degraded image,an image enhancing algorithm based on the generalized histogram is proposed in this paper.When generate the generalized histogram,the count allocated to each pixel is split into two parts.One part is accumlated to the current histogram and the other part is distributed uniformly over the intensity levels.Then the intensity mapping function is generated according to the generalized histogram.For each pixel,only one part is accumulated when generating the generalized histogram,so contrast overenhancement can be alleviated.The experiment results indicate that the proposed method has a better performance than global histogram equalization and local histogram equalization.