【作者】 黄强；

【导师】 傅明； 陈罗生；

【作者基本信息】 长沙理工大学 ， 交通运输， 2007， 硕士

【摘要】 随着现代经济的高速发展，交通运输的保障就显得尤其重要，对交通管理的要求也越来越高，将计算机科学与通信等高新技术运用于交通监控管理与车辆控制，以保障交通顺畅及行车安全，从而改善环境质量，促进经济发展的智能交通系统ITS(Intelligent traffic system，ITS)也随之应运而生。在智能交通管理系统中，实时获取交通车流量的车辆检测技术是ITS的基础。利用数字图像处理技术来实现交通流量的车辆检测技术已成为该研究领域的热点。基于数字图像处理技术的交通流量车辆检测技术的研究始于上个世纪80年代。到现在，检测思想和算法一直在不断的改进和革新。有的研究基于检测区域，有的研究基于整幅图片，大量的文献都是基于整幅图片的研究，如背景差分方法，帧差分方法等。本文分析和比较了常见的车流量检测所涉及到的图像预处理算法和图像识别算法以及车流量检测算法，比如背景差法、帧差法、路面标记法、边缘检测法等，提出了改进的基于边缘信息的车流量检测算法。图像的边缘携带了图像的大部分信息，包含了图像的基本特征。边缘检测是图像处理和模式识别的重要方法，如图像分割和自动目标识别等，是图像处理研究领域的重要课题之一。本文改进后的边缘检测算法结合了改进帧差法与边缘检测法的优点，具有很强的环境自适应能力、计算量小、检测精度高、可靠实用等优点。在本文的结尾，提出了设计中需要改进的地方，需以后对其进行进一步的研究。更多还原

【Abstract】 With the rapid development of the modern economy, keeping traffic and transportation safe and orderly is becoming more and more important, and the demand for traffic management is higher than before. To protect traffic condition, improve transportation environment, and accelerate economic development, Intelligent Traffic System which is based on computer science and communication technology managing traffic and detecting vehicles is created. Technology of the vehicle detection is the key to the series of technologies for the intelligent traffic system. Research on the technology for the vehicle detection based on image processing is just the hot of the field.Research on detecting vehicles based on image processing began on 1980s. Up to now, detecting ideas and detecting technology has been updating. Some researches are based on detecting region, and some are based on the whole picture such as background difference method and frame difference method.A new traffic flow detection algorithm based on edge information with limit progress is proposed in this paper after systematic analysis and comparison of several fashionable detection-related image preprocessing and recognition algorithm and flow detection methods, which background difference method, frame difference method, road mark method and edge detection technology are listed in the consideration. Edges of an image reflect the information of the image mostly including the basic characteristics. Edge detection is a vital part of many image processing and pattern recognition systems. Typical areas of applications are image segmentation, stereo vision, and identification of objects as in automatic target recognition. Edge detection is one of the most important parts in image processing. We believe that this algorithm owns merits of robust adaptation, small calculation; precise and strong stability under certain circumstance for it inherits from both advanced frame difference method and improved edge detection technology.In the end, the limitation has been pointed out as this design and the way to improve. Certainly, we should continue to make research on this topic.更多还原