摘要
为了提高快速定位交通标志的准确性,通过对场景中复杂的直线信息以及颜色信息的融合,提出了一种基于直线包络的快速定位交通标志的方法。该方法首先利用颜色信息以及统计直线信息获取目标候选区域,将交通标志进行粗定位,然后利用形态学运算获取连通区域,再利用交通标志的纵横比和内部孔洞信息将交通标志从目标候选区域中准确定位出来。实验表明该方法可以提高交通标志在夜晚等复杂背景中定位的准确性。
In order to improve the accuracy of fast mation and color information in a scene, a pos fast method uses color information and statistical line traffic signs, through the fusion of the complex linear infor- positioning method based on linear envelope was proposed. The information to obtain the target candidate region for coarse positio- ning of traffic signs, and then, uses morphological operations to obtain regional connectivity, and uses the traffic sign aspect ratio and the void information to position the traffic mental results show that the proposed method can improve the backgrounds like night signs from the target candidate region. The experi- accuracy of the traffic sign positioning in complex
出处
《高技术通讯》
北大核心
2017年第5期450-456,共7页
Chinese High Technology Letters
基金
河北省自然科学基金(F2016203211)
燕大青年教师自主研究计划(15LGA014)资助项目
关键词
机器视觉
连通区域
城市道路
直线包络
单幅图像
交通标志
machine vision, connected region, city roads, linear envelop, single image, traffic sign