基于机器学习算法的交通标志图像智能识别  

Intelligent Recognition of Traffic Sign Image Based on Machine Learning Algorithm

作　　者：曹海燕 张大维 CAO Haiyan;ZHANG Dawei(School of Communication Technology,Nanguang College of Communication University of China,Nanjing 211172,China)

机构地区：[1]中国传媒大学南广学院传媒技术学院,江苏南京211172

出　　处：《微型电脑应用》2021年第1期19-22,共4页Microcomputer Applications

基　　金：江苏省高校自然科学研究面上项目(18KJD520006)。

摘　　要：交通标志图像识别具有十分重要的研究意义,传统方法的交通标志图像识别正确率低,耗时长,为了解决传统方法存在的局限性,更加准确识别各种类型的交通标志图像,提出了基于机器学习算法的交通标志图像智能识别方法。首先对当前交通标志图像识别的研究进展进行分析,找到引起交通标志图像识别误差的影响因素,然后提取多特征交通标志图像特征,并引入机器学习算法——极限学习机拟合特征向量与交通标志图像类型之间的内存关联,建立交通标志图像智能识别的分类器,进行了仿真对照测试。交通标志图像识别正确率超过了95%,将识别误差控制在实际应用的区间内,交通标志图像识别时间短,而且整体识别效果要优于传统方法,验证了交通标志图像智能识别方法的优越性。Traffic sign image recognition has an important research significance.The traditional method of traffic sign image recognition has low accuracy rate,and is time-consuming.In order to solve theproblems of traditional methods,and get more accurate identification of various types of traffic sign images,this paper proposes a traffic sign image intelligent recognition method based on machine learning algorithm.Firstly,this paper analyzes the current research progress of traffic sign image recognition,finds out the influencing factors of traffic sign image recognition error,then extracts multi feature traffic sign image features,and introduces machine learning algorithm extreme learning machine to fit the memory relationship between the feature vector and the traffic sign image type,and establishes the intelligent recognition classifier of traffic sign image.The simulation results show thatthe recognition accuracy of this method is more than 95%,and the recognition error is controlled within the practical application range.The recognition time of traffic sign image is short,and the overall recognition effect is better than the traditional method,which verifies the superiority of the intelligent recognition method of traffic sign image in this paper.

关 键 词：交通标志 图像类型 机器学习算法 智能识别 极限学习机 特征向量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]