基于机器视觉的公路交通标志自动化巡检系统

针对既有交通标志巡检的调查手段和分析方法存在的成本高、效率低、精度低的弊端，提出了一种基于机器视觉的公路交通标志自动化巡检系统，实现对公路交通标志的移动式实时检测与分析。首先利用YOLO （You Only Look Once） 算法，将机器视觉技术运用到交通标志的检测中，加快了检测速度。然后给出了交通标志自动化巡检系统的系统构架、系统结构与功能及数据管理方案。最后利用该系统进行实测。实测结果显示，该自动化巡检系统能够准确地检测交通标志、道路标线及路面坑塘，具有较高的测试精度及较好的测试效果。

Road Traffic Sign Automatic Inspection System Based on Machine Vision

To avoid the drawbacks such as high cost, low efficiency, and low precision of the survey means and analysis methods of traditional traffic sign inspection system, an automatic inspection system was proposed based on machine vision to realize the mobile real-time detection and analysis of highway traffic signs. Firstly, based on the YOLO (You Only Look Once) algorithm, the application of machine vision technology in detecting traffic signs was realized, which speeded up the detection. Secondly, the design of system architecture, system structure and function, and data management scheme of the traffic sign automatic inspection system were given detailedly. Finally, the actual measurement experiment was carried out using the proposed system and the measured results of the system were obtained. The results shows that this automatic inspection system can accurately detect traffic signs, road markings and pavement pits with high accuracy and approving effect.