摘  要

进入二十一世纪之后，网络发展越来越迅速，人们的很多生活与消费习惯都随之发生了改变。在图像识别或深度学习算法不发达的年代，人们是无法通过电脑进行交通标志或者路况和天气的识别的，人们通常是要去各地区的主要道路或者在不同的天气下进行拍照，通过人工的方式进行内容的识别。但这种形式费事费力，因为往往一条道路是有很多交通标志的，比如禁止停车、限速标志或禁止掉头等，每一个城市的工作人员的数量肯定是有限的，所以如果想完成所有交通标志的识别需要很长的周期，并且对工作人员来说强度和难度也非常的大。还有一个原因，不同天气下的路况是不一样的，天气往往是变幻莫测的，所以无法保证工作人员可以在不同的天气下进行素材的手机，不过随着许多传统的行业逐渐与互联网接轨，各种算法及识别技术越来越发达，交通标志等图像的识别也逐渐被网络化的识别系统所替代了。在计算机刚开始发展的时候就出现了许多的图像保存或图像分类系统，但是因为技术的限制导致系统并不完美，有很多不符合使用者使用习惯的瑕疵，也有很多的功能缺陷。随着计算机编程语言的不断发展和移动设备的出现，交通标志等图像识别服务逐渐朝着更专业、更精准、更效率的方向发展。

本系统前台界面使用了最新的HTML5技术，使用DIV+CSS进行布局，使整个前台页面变得更美观，极大的提高了用户的体验，另外本系统无论是使用电脑的浏览器进行访问还是使用移动设备进行访问，都可以保证网站正确的排版。后端的代码技术选择的是PYTHON，PYTHON语言是当下最常用的编程语言之一，可以保证系统的稳定性和流畅性，PYTHON可以灵活的与数据库进行连接。本系统的数据使用的MYSQL数据库，它可以提高查询的速度，增强系统数据存储的稳定性和安全性。本系统的核心算法是深度学习算法，是当下很流行的一种机器学习语言，也是未来很多行业的发展趋势，利用深度学习方法可以准确的识别各类交通标志。

关键词：深度学习；系统；PYTHON；MySQL

Abstract

In the 21st century, with the rapid development of the Internet, many people's lives and consumption habits have changed. In the era when image recognition or depth learning algorithms were not developed, people could not recognize traffic signs or road conditions and weather through computers. People usually wanted to go to the main roads in various regions or take photos in different weather to identify contents manually. However, this form is laborious, because there are often many traffic signs on a road, such as no parking, speed limit signs or no turning around. The number of staff in each city is certainly limited. Therefore, it takes a long period to complete the identification of all traffic signs, and the intensity and difficulty for staff are also very large. Another reason is that the road conditions are different in different weather conditions, and the weather is often unpredictable, so it is impossible to guarantee that staff can use mobile phones in different weather conditions. However, as many traditional industries gradually connect with the Internet, various algorithms and recognition technologies are increasingly developed, and the recognition of traffic signs and other images is gradually replaced by networked recognition systems. At the beginning of computer development, many image storage or image classification systems appeared. However, due to technical limitations, the system is not perfect. There are many flaws that do not conform to the user's usage habits, and there are also many functional defects. With the continuous development of computer programming languages and the emergence of mobile devices, image recognition services such as traffic signs are gradually becoming more professional, more accurate and more efficient.

The foreground interface of this system uses the latest HTML5 technology and uses DIV+CSS for layout, which makes the whole foreground page more beautiful and greatly improves the user experience. In addition, this system can ensure the correct layout of the website no matter whether it is accessed using a computer browser or a mobile device. The back-end code technology is PYTHON. PYTHON language is one of the most commonly used programming languages, which can ensure the stability and smoothness of the system. PYTHON can flexibly connect with the database. The data of this system uses the MYSQL database, which can improve the speed of query and enhance the stability and security of system data storage. The core algorithm of this system is the deep learning algorithm, which is a popular machine learning language at present, and also the development trend of many industries in the future. The deep learning method can accurately identify various traffic signs.

Key words: Deep learning; System; PYTHON； MySQL