CAJ | 학술논문

交通灯检测是智能交通系统研究的一个重要方向，采用了一种交通灯实时检测算法。首先，对图像进行RGB归一化处理，得到了几种颜色独立的图像。然后提取图像的二值化信息，并作形态学处理。接着对图像进行分群操作和标记处理，并且根据交通灯形状区域特点做出边长与面积条件判断。最后通过提取交通灯的样本空间来进行匹配，从而锁定交通灯位置。实验部分将其与传统的 Hough 算法进行比较，结果表明所提算法可以更快且准确地检测到交通灯。
교통등검측시지능교통계통연구적일개중요방향，채용료일충교통등실시검측산법。수선，대도상진행RGB귀일화처리，득도료궤충안색독립적도상。연후제취도상적이치화신식，병작형태학처리。접착대도상진행분군조작화표기처리，병차근거교통등형상구역특점주출변장여면적조건판단。최후통과제취교통등적양본공간래진행필배，종이쇄정교통등위치。실험부분장기여전통적 Hough 산법진행비교，결과표명소제산법가이경쾌차준학지검측도교통등。
Traffic lights detection is an important research direction of intelligent transportation system with adopts the real-time detection algorithm adopted.Firstly,the image is processed by RGB normali-zation model,which produces several independent image colors.Secondly,the algorithm extracts the bina-rization information from the image and makes the morphological processing.Thirdly,it exerts the clus-tering operation and labeling processing for each image and makes the condition j udgment on the length and area according to the shape characteristics of traffic lights region.Finally,the value is extracted from the sample space of traffic lights and matched with the one from the real-time scene.Therefore,the posi-tion of traffic lights is locked well.The experiment compares the Clustering algorithm with the tradition-al Hough algorithm and the results show that the proposed algorithm can quickly and accurately detect the traffic lights.