基于亮度对比度和暗原色先验原理的白天道路图像能见度检测方法
|
摘要
低能见度是对道路通行影响最为严重的气象要素之一。随着数字摄像技术和图像识别技术的发展以及气象和交通部门间信息共享工作的开展,利用高速公路沿线摄像头视频数据快速识别能见度成为提高能见度时空监测精度的重要手段。本文提出了一种基于亮度对比度和暗原色先验原理的白天道路图像能见度检测方法。首先根据霍夫变换直线检测方法提取道路兴趣域,然后根据亮度对比度方法检测人眼可分辨最远像素点,将其作为目标点,最后基于暗原色先验原理求取目标点的透射率,并根据能见度与消光系数的关系公式求取图像能见度值。利用安徽省京台高速吴玗北段和宁绩高速宁国互通段视频图像资料和邻近交通气象站能见度监测资料,采用绝对误差和能见度等级误差对能见度检测效果进行检验。结果表明,本方法对能见度的变化较为敏感,能见度等级的检测效果较好,准确度可达95%,对开展公路交通视频图像能见度识别工作具有较好借鉴应用意义。
Low visibility is one of the most serious meteorological factors affecting road traffic.Along with the technology development of digital camera and image recognition and the information shared among the meteorological and traffic departments,using camera video data along the expressway to rapid recognize the visibility becomes an important method to improve the precision of visibility temporal and spatial monitoring.In this paper,we propose a daytime road image visibility detection method based on the brightness contrast and dark channel prior principle:(1)to extract the interested road domain according to the Hough transform detection method,(2)by using the brightness contrast method,the farthest pixel point that the human eye can distinguish is used as the target point,(3)to evaluate the transmittance of the target point based on the dark channel prior principle,(4)to figure out the image visibility value by the relationship between visibility and extinction coefficient.The test indicators of this paper are visibility absolute error and visibility level error.The test data is derived from the video image data of the Wuyu section of Jingtai Expressway in Anhui Province and the Ningguo-Hutong section of Ningji Expressway in Anhui province,and the visibility data set of the traffic weather stations near the expressway in Anhui Province.The results indicate that the method can accurately reflect the trend of visibility change,and the detection accuracy can be as high as 95%,which can be a good reference for the image visibility recognition.
Low visibility is one of the most serious meteorological factors affecting road traffic.Along with the technology development of digital camera and image recognition and the information shared among the meteorological and traffic departments,using camera video data along the expressway to rapid recognize the visibility becomes an important method to improve the precision of visibility temporal and spatial monitoring.In this paper,we propose a daytime road image visibility detection method based on the brightness contrast and dark channel prior principle:(1)to extract the interested road domain according to the Hough transform detection method,(2)by using the brightness contrast method,the farthest pixel point that the human eye can distinguish is used as the target point,(3)to evaluate the transmittance of the target point based on the dark channel prior principle,(4)to figure out the image visibility value by the relationship between visibility and extinction coefficient.The test indicators of this paper are visibility absolute error and visibility level error.The test data is derived from the video image data of the Wuyu section of Jingtai Expressway in Anhui Province and the Ningguo-Hutong section of Ningji Expressway in Anhui province,and the visibility data set of the traffic weather stations near the expressway in Anhui Province.The results indicate that the method can accurately reflect the trend of visibility change,and the detection accuracy can be as high as 95%,which can be a good reference for the image visibility recognition.
引文
[1]袁成松,卞光辉,冯民学,等.高速公路上低能见度的监测与预报[J].气象,2003,29(11):36-40.
[2]田小毅,袁成松,吴震.高速公路上霾雾演变及其对能见度的影响[J].气象科技,2010,38(6):673-678.
[3]周须文,时青格,贾俊姝,等.低能见度雾的分级预报方法研究[J].热带气象学报,2014,30(1):161-166.
[4]蒋璐璐,魏鸣.FY-3A卫星资料在雾监测中的应用研究[J].遥感技术与应用,2011,26(4):489-495.
[5]张顺谦,杨秀蓉.基于神经网络和分形纹理的夜间浓雾遥感监测技术[J].应用气象学报,2005,16(6):804-810.
[6]张伟康,马慧云,邹峥嵘,等.基于SBDART和BP的夜间雾遥感检测和能见度反演[J].解放军理工大学学报(自然科学版),2014,15(2):197-202.
[7]傅刚,徐杰,张树钦.数值模拟和卫星反演大气能见度对比分析[J].中国海洋大学学报(自然科学版),2011,41(4):1-10.
[8]李永平,刘晓波,葛伟强,等.一种基于卫星遥感和数值预报产品的雾预报方法[J].应用气象学报,2012,23(3):340-347.
[9]Kwon T M.Atmospheric visibility measurements using video cameras:Relative visibility[R].America:University of Minnesota Duluth,2004.
[10]Hallowell R G,Matthews M P,Pisano P A.Automated extraction of weather variables from camera imagery[C]∥Proceedings of the 2005 Mid-Continent Transportation Research Symposium,Ames,IA,2005:1-13.
[11]Final report on signal and image processing for road condition classification[R].Stockholm:Swedish National Road Agency,2002.
[12]李佳,葛嘉琦,陈启美.路况视频能见度检测算法[J].计算机工程,2009,35(16):175-177.
[13]吕伟涛,阳善昌,刘亦风,等.基于数字摄像技术测量气象能见度---双亮度差方法和试验研究[J].大气科学,2004,28(4):559-568.
[14]陈钊正,陈启美.基于摄像机自标定的视频对比度能见度检测算法与实现[J].电子与信息学报,2010,32(12):2907-2912.
[15]李勃,董蓉,陈启美.无需人工标记的视频对比度道路能见度检测[J].计算机辅助设计与图形学学报,2009,21(11):1575-1582.
[16]张潇,李勃,陈启美.基于亮度特征的PTZ视频能见度检测算法及实现[J].仪器仪表学报,2011,32(11):381-386.
[17]胡平,杨旭东.高速公路能见度快速检测算法[J].公路交通科技,2017,34(4):115-122.
[18]郭尚书,齐文新,齐宇.基于暗通道先验的视频能见度测量方法[J].计算机与数字工程,2014,42(4):694-697.
[19]吴炜,李勃,杨娴,等.基于路面视亮度差平方最优化的视频能见度检测算法[J].电子与信息学报,2014,36(10):2413-2418.
[20]李晓航,郭佳,彭富伦,等.一种改进的基于Hough变换的道路图像检测方法[J].应用光学,2016,37(2):229-234.
[21]He Kaiming,Sun Jian,Tang Xiaoou.Single image haze removal using dark channel prior[C]∥Proceedings of IEEEConference on Computer Vision and Pattern Recognition(CVPR).New York:IEEE,2009:1956-1963.
[22]QX/T 111-2010高速公路交通气象条件等级[S].
[2]田小毅,袁成松,吴震.高速公路上霾雾演变及其对能见度的影响[J].气象科技,2010,38(6):673-678.
[3]周须文,时青格,贾俊姝,等.低能见度雾的分级预报方法研究[J].热带气象学报,2014,30(1):161-166.
[4]蒋璐璐,魏鸣.FY-3A卫星资料在雾监测中的应用研究[J].遥感技术与应用,2011,26(4):489-495.
[5]张顺谦,杨秀蓉.基于神经网络和分形纹理的夜间浓雾遥感监测技术[J].应用气象学报,2005,16(6):804-810.
[6]张伟康,马慧云,邹峥嵘,等.基于SBDART和BP的夜间雾遥感检测和能见度反演[J].解放军理工大学学报(自然科学版),2014,15(2):197-202.
[7]傅刚,徐杰,张树钦.数值模拟和卫星反演大气能见度对比分析[J].中国海洋大学学报(自然科学版),2011,41(4):1-10.
[8]李永平,刘晓波,葛伟强,等.一种基于卫星遥感和数值预报产品的雾预报方法[J].应用气象学报,2012,23(3):340-347.
[9]Kwon T M.Atmospheric visibility measurements using video cameras:Relative visibility[R].America:University of Minnesota Duluth,2004.
[10]Hallowell R G,Matthews M P,Pisano P A.Automated extraction of weather variables from camera imagery[C]∥Proceedings of the 2005 Mid-Continent Transportation Research Symposium,Ames,IA,2005:1-13.
[11]Final report on signal and image processing for road condition classification[R].Stockholm:Swedish National Road Agency,2002.
[12]李佳,葛嘉琦,陈启美.路况视频能见度检测算法[J].计算机工程,2009,35(16):175-177.
[13]吕伟涛,阳善昌,刘亦风,等.基于数字摄像技术测量气象能见度---双亮度差方法和试验研究[J].大气科学,2004,28(4):559-568.
[14]陈钊正,陈启美.基于摄像机自标定的视频对比度能见度检测算法与实现[J].电子与信息学报,2010,32(12):2907-2912.
[15]李勃,董蓉,陈启美.无需人工标记的视频对比度道路能见度检测[J].计算机辅助设计与图形学学报,2009,21(11):1575-1582.
[16]张潇,李勃,陈启美.基于亮度特征的PTZ视频能见度检测算法及实现[J].仪器仪表学报,2011,32(11):381-386.
[17]胡平,杨旭东.高速公路能见度快速检测算法[J].公路交通科技,2017,34(4):115-122.
[18]郭尚书,齐文新,齐宇.基于暗通道先验的视频能见度测量方法[J].计算机与数字工程,2014,42(4):694-697.
[19]吴炜,李勃,杨娴,等.基于路面视亮度差平方最优化的视频能见度检测算法[J].电子与信息学报,2014,36(10):2413-2418.
[20]李晓航,郭佳,彭富伦,等.一种改进的基于Hough变换的道路图像检测方法[J].应用光学,2016,37(2):229-234.
[21]He Kaiming,Sun Jian,Tang Xiaoou.Single image haze removal using dark channel prior[C]∥Proceedings of IEEEConference on Computer Vision and Pattern Recognition(CVPR).New York:IEEE,2009:1956-1963.
[22]QX/T 111-2010高速公路交通气象条件等级[S].