\r董 娜，刘欣宇，吴爱国\r
\r
AuthorsHTML:\r董 娜，刘欣宇，吴爱国\r
\r
AuthorsListE:\rDong Na，Liu Xinyu，Wu Aiguo\r
\r
AuthorsHTMLE:\rDong Na，Liu Xinyu，Wu Aiguo\r
\r
Unit:\r天津大学电气自动化与信息工程学院，天津 300072\r
\r
Unit_EngLish:\rSchool of Electrical and Information Engineering，Tianjin University，Tianjin 300072，China\r
\r
Abstract_Chinese:\r道路交通标志检测是智能交通系统的重要组成部分，精确快速识别圆形交通标志对保障交通安全有非常关键的作用．本文将子种群概念引入人工蜂群(artificial bee colony，ABC)算法中，提出了一种基于子种群多峰值优化算法(species-based artificial bee colony，SABC)应用于圆形交通标志的检测．SABC的关键在于多个子种群按照人工蜂群算法同时搜索多个最优解．作为比较，引入了基于子种群的遗传算法(species-based genetic algorithm，SGA)和基于子种群的粒子群优化算法(species-based particle swarm optimization，SPSO)．5 个多峰值函数用于验证3 种算法的多峰值优化性能，其结果证明SABC 具有更高的准确性和更快的运行速度，成功率为100%，精度均低于10-4，运行时间均在0.3 s 以内．然后，将多圆检测视为多峰值优化问题，并将SABC 应用于多圆检测，设计了完整的检测算法．最后，在德国交通标志数据集(German traffic sign recognition benchmarks，GTSRB)的圆形交通标志上测试所提出的圆形检测算法．实验结果表明，所提算法能够以良好的性能定位圆形交通标志．\r
\r
Abstract_English:\rThe detection of road traffic signs is an important component of the intelligent transportation system. The accurate and rapid recognition of circular traffic signs plays a crucial role in ensuring traffic safety. In this paper，we introduce the concept of species to the artificial bee colony(ABC)algorithm for the detection of circular traffic signs，and propose a multi-peak optimization algorithm based on a species-based artificial bee colony(SABC). The key idea of the SABC is that multiple species simultaneously search for multiple optimal values specific to each artificial bee colony. We compare the performance of the SABC with those of the species-based genetic algorithm(SGA)and particle-based particle swarm optimization(SPSO)algorithm. Using five multi-peak functions to verify the multi-peak optimization performance of these three algorithms，the results indicate that SABC has higher accuracy and faster running speed. With a success rate of 100%，SABC’s accuracy is within 10^-4，and its running time is less than 0.3 s. Then，considering multi-circle detection as a multi-peak optimization problem，we applied the SABC to multi-circle detection，and designed a complete detection method. We tested the proposed circular detection method on circular traffic signs with respect to the German Traffic Sign Recognition Benchmark(GTSRB). The experimental results show that the proposed algorithm performs well in locating circular traffic signs.\r
\r
Keyword_Chinese:人工蜂群算法；多峰值优化；子种群；圆形检测；交通标志检测\r

Keywords_English:artificial bee colony(ABC)algorithm；multi-peak optimization；species；circle detection；traffic signs detection\r


PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=6242