一种基于亮度直方图的自动曝光控制方法

自动曝光控制是现代相机和摄像机必备功能之一。提出一种基于亮度直方图的自动曝光控制算法。基于人类视觉系统(HVS)的视觉注意机制,可认为直方图中大而陡峭的峰值区域对应于图像中的不感兴趣区域。算法寻找直方图中最大的两个峰值区域,并依据峰值区域的大小确定它们中像素亮度的加权值,从而计算图像的加权亮度均值。像素亮度的加权值与其所属峰值区域的大小之间的关系由一组二次型曲线描述,二次型曲线的参量是由图像的背景亮度决定,其中背景亮度由自动曝光系统参量计算得到。在实际应用中采用了模糊逻辑确定最大的两个峰值区域的加权值。实验结果表明该算法对各种场景均能进行有效的自动曝光控制。

A New Automatic Exposure Algorithm for Video Cameras Using Luminance Histogram

Automatic exposure control(AEC) is one of the indispensable functions of modern digital still cameras and video cameras.A new automatic exposure algorithm using luminance histogram of an image is proposed.According to the attention mechanism of human visual system(HVS),it is considered that peak regions in luminance histogram correspond to the regions of no interest in an image.The aim of the algorithm is to find the two largest peak regions in histogram and calculate the mean weighted luminance of entire image.The luminance weights of pixels in peak regions are decided by a set of quadratic curves,and the parameters of quadratic curves are affected by the brightness of image background,which can be obtained from parameters of the automatic exposure system.Fuzzy logic is applied to practical systems to determine the luminance weights of pixels in the two largest peak regions.Results show that the new automatic exposure algorithm gives efficient exposure control over various scene tests.