Abstract:

Traffic sign detection and recognition is one of the research hotspots of environment perception in three major modules of unmanned driving. Detection and recognition of traffic signs can transmit road traffic information to unmanned vehicles and optimize driving decisions. In complex environments such as heavy rain, heavy fog and low light, the captured images are often blocked and blurred. This not only affects the quality of the image, but also brings huge difficulties to the detection and recognition of the later signs. Therefore, this paper descripes briefly the traffic sign detection and recognition methods, and summarizes the methods, principles and steps of domestic and foreign scholars in recent years to solve traffic sign detection and recognition in various complex environments, which are conducive to better solve such problems latter. At the same time, the commonly used traffic sign data set is summarized, and the proportion of images taken in the complex environment in the data set is explained.

Key words: traffic sign detection and recognition, complex environment, traffic sign data set

摘要：

交通标志检测与识别是无人驾驶三大模块中环境感知的研究热点之一，检测和识别交通标志可以向无人车传递道路交通信息，优化行车决策。在暴雨、大雾以及光线昏暗等复杂环境下，拍摄到的图像往往会被遮挡，变得模糊。这不仅影响图像的质量，还会对后期标志的检测与识别带来巨大的困难。简述了交通标志检测与识别方法，对近年来国内外学者解决各类复杂环境下交通标志检测与识别的方法、原理和步骤进行了总结归纳，有利于人们更好地解决此类问题。同时，对常用的交通标志数据集进行了总结，并对数据集里在复杂环境下拍摄的图像比例给予了说明。

关键词: 交通标志检测与识别, 复杂环境, 交通标志数据集