弱光线下道路交通标志图像增强算法  

Road traffic sign image enhancement algorithm under low light conditions

作　　者：张书涵 王婷 郭钊汝 Zhang Shuhan;Wang Ting;Guo Zhaoru(College of Computer and Information Engineering,Xinjiang Agricultural University,Urumqi,Xinjiang 830052,China)

机构地区：[1]新疆农业大学计算机与信息工程学院,新疆乌鲁木齐830052

出　　处：《计算机时代》2025年第1期5-10,共6页Computer Era

摘　　要：针对弱光线下道路交通标志特征不明显、细节颜色丢失,目标检测算法识别准确度低的问题,本文基于HSV空间提出针对V分量的改进MSR-CLAHE算法、针对S分量的UM-GF算法。对V分量利用双边滤波代替高斯滤波改进MSR算法,进行局部对比度增强,并结合CLAHE算法以限制对比度的过度放大。对S分量采用非锐化掩模图像进行引导滤波,保持边缘锐化,同时也有效保留了色彩信息。并在CCTSDB数据集中选取弱光线图片进行了测试实验,实验的主观和客观评价指标表明本文提出的算法能够有效提高图像对比度,增强细节信息,为后续交通标志识别提供支持。In order to solve the problem that the characteristics of road traffic signs are not obvious and the details are lost under weak light,and the recognition accuracy of target detection algorithm is low,this paper proposes an improved MSR-CLAHE algorithm for the V-component and a UM-GF algorithm for the S-component based on the HSV space.For the V component,the MSR algorithm is improved by using bilateral filtering instead of Gaussian filtering,and the local contrast is enhanced by combining CLAHE algorithm to limit the excessive amplification of contrast.For S component,the unsharpened mask image is used to guide the filtering,which keeps the edge sharpening,and also effectively retains the color information.Low-light images were selected in the CCTSDB dataset for testing experiments.The experiments show that the algorithm adopted in this paper can effectively improve image contrast and enhance details in both subjective and objective evaluation tests,and provide support for subsequent traffic sign recognition.

关 键 词：图像增强 MSR算法 CLAHE算法 非锐化掩模 引导滤波 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]