摘要
介绍一种基于CMOS有源像素传感器为感光元件的宽照度范围的静态图像采集方法.采用非线性曝光二值化图像叠加技术,达到有效降低图像的饱和溢出和改善图像暗场空缺的目的;利用CMOS图像传感器输出像元信号电压与曝光时间的线性响应关系,给出一种非线性曝光时间计算方法,使像元输出电压随光照度变化在一定范围内满足对数响应关系.实验分别采用白色背景下的黑色条带图像采样、室内灯光场景的景物图像采样和室外强光直射场景的仿车牌信息图像采样.结果表明,该方法获得的图像偏差率仅为传统CMOS相机探测的图像偏差率的1/7,实现了光照度在60 dB内变化图像信息的清晰识别,可应用于宽照度下车辆牌照等静态图像采集.
In this paper,a static image acquisition method with wide illumination range based on CMOS active pixel sensor is introduced.The nonlinear exposure binary image superposition technology is used to effectively reduce the saturation overflow of the image and improve the vacancy of the dark field of the image.Based on the linear response relationship between the output pixel signal voltage of CMOS image sensor and the exposure time,a nonlinear exposure time calculation method is given to make the pixel output voltage meet the logarithmic response relationship in a certain range with the change of illumination.The experiment adopts black strip image sampling in white background,scene image sampling in indoor lighting scene and anti-license plate image sampling in outdoor strong light direct scene.The experimental results show that the image deviation rate obtained by the proposed method is only 1/7 of the image deviation rate detected by traditional CMOS camera.It realizes the clear recognition of the image with the change of illuminance in the range of 60 dB,and can be applied to the static image acquisition of vehicle license plate under wide illuminance.
作者
黄梅红
潘娇
郑齐清
戴官全
吴平辉
吴豪杰
HUANG Meihong;PAN Jiao;ZHENG Qiqing;DAI Guanquan;WU Pinghui;WU Haojie(School of Transportation and Navigation,Quanzhou Normal University,Quanzhou Fujian 362000,China;Key Laboratory of Information Functional Material of Fujian Higher Education,School of Physics and Information Engineering,Quanzhou Normal University,Quanzhou Fujian 362000,China;Faculty of Mathematics and Computer Science,Quanzhou Normal University,Quanzhou Fujian 362000,China)
出处
《泉州师范学院学报》
2023年第2期9-15,70,共8页
Journal of Quanzhou Normal University
基金
福建省教育厅项目(JAT190534,JAT200563)
泉州师范学院大学生创新创业训练计划资助项目(202110399083)。
关键词
静态图像采集
CMOS传感器
二值化
图像叠加
宽照度
图像偏差率
static image acquisition
CMOS sensor
binarization
image superposition
wide illumination
image deviation rate
