| BP神经网络下的限速交通标志实时检测识别 |
| |
| 引用本文： | 张兴国,刘晓磊,李靖,王环东. BP神经网络下的限速交通标志实时检测识别[J]. 西安电子科技大学学报(自然科学版), 2018, 45(5): 136-142. DOI: 10.3969/j.issn.1001-2400.2018.05.022 |
| |
| 作者姓名： | 张兴国 刘晓磊 李靖 王环东 |
| |
| 作者单位： | (1. 中国飞行试验研究院，陕西 西安 710089；2. 武汉大学 遥感信息工程学院，湖北 武汉 430072) |
| |
| 基金项目： | 国家自然科学基金资助项目(F050105) |
| |
| 摘 要： | 目前已有很多基于交通标志图片进行道路交通标志识别的系统研究，但存在误识别率较高的问题．因此文中基于视频数据对限速交通标志进行检测与识别，实现了限速交通标志的自动检测定位，并采用反向传播神经网络来进行道路标志的识别; 同时，还采用连续自适应的均值漂移算法和光流法进行视频加速．实验表明，新提出的算法在耗时上缩短50%以上，检测识别准确度在90%以上，适用于相关智能识别领域．
|
| 关 键 词： | 限速交通标志 自动检测定位 人工神经网络 视频加速 |
| 收稿时间： | 2017-11-06 |
Real-time detection and identification of speed limit traffic signs under the BP neural network |
| |
| ZHANG Xingguo,LIU Xiaolei,LI Jing,WANG Huandong. Real-time detection and identification of speed limit traffic signs under the BP neural network[J]. Journal of Xidian University, 2018, 45(5): 136-142. DOI: 10.3969/j.issn.1001-2400.2018.05.022 |
| |
| Authors: | ZHANG Xingguo LIU Xiaolei LI Jing WANG Huandong |
| |
| Affiliation: | (1. Chinese Flight Test Establishment, Xi'an 710089, China;2. School of Remote Sensing and Information Engineering, Wuhan Univ., Wuhan 430072, China) |
| |
| Abstract: | There are a lot of traffic signs based on the picture system of traffic sign recognition, and based on the video data to detect and identify the speed of traffic signs, However, there is a higher error rate. So this paper realizes the automatic detection and localization of speed limit traffic signs, and uses the BP neural network to identify the road signs. Meanwhile, the CamShift method and the optical flow method are used to speed up video, Experiment shows that the algorithm proposed in this paper can shorten the time by more than 50%, and the detection and recognition accuracy is over 90%, which is suitable for the related field of intelligent recognition. |
| |
| Keywords: | traffic signs of speed limit automatic detection and location propaqation artificial neural network video acceleration |
|
| 点击此处可从《西安电子科技大学学报(自然科学版)》浏览原始摘要信息 |
|
点击此处可从《西安电子科技大学学报(自然科学版)》下载全文 |
