基于加权灰度熵差的分区域曝光算法  

REGIONAL EXPOSURE ALGORITHM BASED ON WEIGHTED D-VALVE OF GRAY ENTROPY

作　　者：胡晓彤[1] 刘楠 朱博文 Hu Xiaotong;Liu Nan;Zhu Bowen(College of Computer Science and Information Engineering,Tianjin University of Science and Technology,Tianjin 300457,China)

机构地区：[1]天津科技大学计算机科学与信息工程学院,天津300457

出　　处：《计算机应用与软件》2020年第12期177-183,共7页Computer Applications and Software

基　　金：国家自然科学基金项目(61502339)。

摘　　要：针对传统曝光算法在景物对比度差异较大的监控场景下难以得到均衡曝光的问题,提出基于加权灰度熵差的分区域曝光算法。使用融合了灰度均值与灰度值概率分布信息的加权灰度熵差作为曝光控制参数,增强了曝光控制参数的场景描述能力。为解决单一曝光参数无法得到均衡曝光的问题,通过曝光时间估算、分区曝光和图像拼接完成分区域曝光成像。实验表明,该算法在复杂场景下也能够得到较好的曝光效果,提高了监控摄像头等成像设备的环境适应能力。It is difficult to get the actual exposure in monitoring scenes with large contrast difference by traditional exposure algorithm.This paper proposes a regional exposure algorithm based on weighted D-valve of gray entropy,which integrates the average gray level and distribution of gray-level value,and it enhances the ability of image description of the exposure control parameter.As a single exposure parameter can’t get the balanced exposure,the algorithm completes the sub-area exposure imaging by estimating exposure time,sub-area exposure and image mosaic.The experiments show that the proposed algorithm can get better exposure effect in complex scenes and improve the environmental adaptability of imaging equipment as surveillance cameras.

关 键 词：曝光算法 加权灰度熵差 曝光估算 分区域曝光 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]