|
|||||
|
|
| 基于字袋模型的交通标志识别方法研究 | |
| 引用本文： | 胡晓光,朱欣焰,柳林,李卉,李德仁.基于字袋模型的交通标志识别方法研究[J].测绘科学,2012,37(6):107-110. |
| 作者姓名： | 胡晓光 朱欣焰 柳林 李卉 李德仁 |
| 作者单位： | 1. 武汉大学测绘遥感信息工程国家重点实验室,武汉,430072 2. 武汉大学测绘遥感信息工程国家重点实验室,武汉430072;山东科技大学测绘学院,山东青岛266510 3. 中国地质大学,武汉,430074 |
| 基金项目： | 武汉大学测绘遥感信息工程国家重点实验室自主探索课题资助项目，武汉大学测绘遥感信息工程国家重点实验室专项基金资助项目，武汉大学测绘遥感信息工程国家重点实验室开放研究基金资助项目 |
| 摘 要： | 交通标志识别是智能交通中一个重要的研究领域,实现辅助驾驶和自动导航等应用正获得越来越多的关注.本文提出一种自然场景下交通标志的识别方法:提取标志的局部特征以提高获取效率也使其较少受到遮挡的影响,通过字袋模型量化后获得标志的本征特征,基于该本征特征训练支持向量机,最后使用得到的判别分类器进行交通标志的分类.实验结果表明,... |
| 关 键 词： | 交通标志识别 字袋模型 本征特征 支持向量机 |
Recognition of traffic signs based on Bag-of-Words model |
|
| HU Xiao-guang , ZHU Xin-yan , LIU Lin , LI Hui , LI Deren.Recognition of traffic signs based on Bag-of-Words model[J].Science of Surveying and Mapping,2012,37(6):107-110. | |
| Authors: | HU Xiao-guang ZHU Xin-yan LIU Lin LI Hui LI Deren |
| Institution: | ①(①State key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan 430072,China;②Shandong University of Science and Technology,Shandong Qingdao 266510,China;③China University of Geosciences,Wuhan 430074,China) |
| Abstract: | TSR(Traffic sign recognition)is an important studied field in ITS(Intelligent traffic system)and is gaining more and more attention for realizing drivers assisting system and automated navigation etc.The authors proposed a recognition method of traffic sign in natural environments.They acquired the local features that could not only improve the efficiency but also be less affected by occlusion,and quantified these features to acquire the intrinsic features of the sign by bag of words model.SVM(Support vector machine)was trained using these intrinsic features and then got a discriminative classification model to classify different traffic signs.As shown by the experiment results,compared with existing method,the proposed method was prore to improve the recognition rate and robustness. |
| Keywords: | traffic sign recognition bag of words Intrinsic feature SVM |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |