With the development of intelligent transportation technologies, road traffic video surveillance has become a hot spot of the intelligent transportation research. However, since the illumination at night is not enough, the visual quality of the image cannot be satisfied, which makes the traffic safety analysis difficult. Considering the problem on the traffic video image when the illumination is not enough at night, under the support of National Traffic Safety Technology Plan, Topic 1 (Regional Road Network Traffic Safety Information Collection Key Technologies and Equipment) of Project 6 (Regional Road Network Traffic Safety Posture Monitoring, Assessment and Emergency Command), made a detailed study. It has been shown that although the common histogram transformation method can enhance the visual quality of the image under the condition of insufficient illumination, the result of visual enhancement cannot be satisfied, resulting in the problem of false edges and overexposure, making the subsequent image processing difficult, including vehicle and pedestrian recognition. Therefore, this paper proposes an improved method for histogram transformation, aiming to highly improve the visual quality of the traffic video image when the illumination is not enough at night, thus strengthening the management of road traffic safety.%随着智能交通技术的发展,道路交通视频监控成为智能交通研究的热点,但是由于夜间光照不够,监控图像的视觉质量往往令人不满意,不利于交通安全分析.针对夜间照度不够条件下的交通视频图像存在的问题,在国家交通安全科技行动计划的支持下,课题六(区域公路网交通安全态势监测、评估及应急指挥)专题一(区域公路网交通安全信息采集关键技术及设备)对其作了研究.尽管常用的直方图变换方法能提高夜间照度不够条件下图像的视觉质量,但是常常存在伪边缘和过曝光等问题,不利于车牌识别、行人特征识别等后续图像处理.文中提出了改进的直方图变换方法,提高夜间照度不够条件下交通视频图像的视觉质量,加强道路交通安全的管理.
展开▼
