摘要
交通环境日趋复杂,车辆识别技术的应用场景也愈发广泛,通过聚类分析方法对传统ORB算法进行改进,应用于前方车辆识别系统。将ORB提取出的特征向量进行归类及筛选,减少噪点及特征冗余点数量,从而提高匹配效率,经实验对比验证,基于K-means的ORB算法一定程度消除无关物体的干扰。不仅保留了原ORB算法运算快的优越性,且对常用数据库识别准确率提高了约10个百分点。相比于同类算法,有效性和稳定性大幅增加,改进算法在复杂环境下的智慧交通领域具有更好的应用效果和适用性。
The traffic environment is becoming more and more complex,and the application scenarios of vehicle identification technology are becoming more and more extensive.In this paper,the traditional orb algorithm is improved by clustering analysis method and applied to the front vehicle recognition system.The feature vectors extracted from ORB are classified and filtered to reduce the number of noise points and feature redundancy points,so as to improve the matching efficiency.The experimental results show that the orb algorithm based on K-means eliminates the interference of irrelevant objects to a certain extent.It not only retains the advantages of fast operation of the original ORB algorithm,but also improves the recognition accuracy of common databases by about 10 percentage points,compared with similar algorithms The improved algorithm has better application effect and applicability in the field of intelligent transportation in complex environment.
作者
杨昊瑜
戴华林
王丽
张蕊
YANG Haoyu;DAI Hualin;WANG Li;ZHANG Rui(School of Computer and Information Engineering,Tianjin Urban Construction University,Tianjin 300384,China;School of Automobile and Transportation,Tianjin Polytechnic Normal University,Tianjin 300222,China)
出处
《传感器与微系统》
CSCD
2020年第10期157-160,共4页
Transducer and Microsystem Technologies
基金
国家重点研发计划资助项目(2019YFD1100402)
天津市科技支撑重点研发计划资助项目(18YFZCGX00380)
天津市教委科研计划项目(2016CJ12,2018KJ172)。
关键词
聚类分析
ORB算法
K均值
目标检测
智慧交通
clustering analysis
ORB algorithm
K-means
target detection
intelligent transportation