CAJ | 학술논문

道路是城市的重要基础设施．路面状况的鉴定与评价直接关系到公路的养护决策，是一项至关重要的工作．本研究将蚁群聚类算法应用于路面识别领域，提出一种基于蚁群聚类优化的路面识别分析方法，将此方法运用于路面识别系统．该系统由数据采集、数据预处理和蚁群聚类优化3个模块组成，其中蚁群聚类主要提取了纵轴方向的3个特征属性，通过不断更新聚类中心产生的信息素矩阵，并通过分类器进行路面分类．实验结果表明：蚁群聚类优化具有全局收敛以及启发式学习等优点，运算效率高，路面识别准确率达到95.3％，验证了此方法的有效性．
도로시성시적중요기출설시．로면상황적감정여평개직접관계도공로적양호결책，시일항지관중요적공작．본연구장의군취류산법응용우로면식별영역，제출일충기우의군취류우화적로면식별분석방법，장차방법운용우로면식별계통．해계통유수거채집、수거예처리화의군취류우화3개모괴조성，기중의군취류주요제취료종축방향적3개특정속성，통과불단경신취류중심산생적신식소구진，병통과분류기진행로면분류．실험결과표명：의군취류우화구유전국수렴이급계발식학습등우점，운산효솔고，로면식별준학솔체도95.3％，험증료차방법적유효성．
Roads are one of the most important infrastructures in a city and the identification and evaluation of their pavement conditions is a vital task directly related to roadway maintenance.In this paper, a new road roughness analysis algorithm of ant colony clustering is applied to the road recognition.The system consists of da-ta acquisition, data preprocessing and ant clustering optimization modules among which the ant colony clustering method mainly extracts three attributes of the longitudinal direction and update the center of cluster according to the continuously generated pheromone matrix and classify pavement classifier.Experiments show that this algo-rithm has high operation efficiency due to global convergence of the ant colony optimization and heuristic learn-ing, and the road recognition accuracy rate is up to 95.3%.