面向智能交通的单目视觉测距方法研究

与前方车辆距离是影响行车安全的重要因素，因此本文提出一种面向未来智能交通的前方车辆单目视觉测距方法.首先，提出融合物联网、智能识别、云计算技术的车联网模型，车辆可实时向车联网回传位置信息及前车图像，请求附近交通标志及前方车辆几何尺度信息，车辆端可计算图像坐标系下车道标志线、交通标志、车辆尺度信息.然后，建立单目相机数学模型，介绍以交通标志、车道分界线为合作标志的单目视觉测距方法.最后，综合应用单目视觉测距方法，设计了前方车辆自适应视觉测距方案.通过仿真实验，证明了单目视觉测距方法的正确性与有效性，可丰富驾驶辅助系统的前方车辆测距手段.

High Precision Distance Measurement Based on Monocular Vision for Intelligent Traffic

The preceding vehicle distance is a significant factor, affecting driving safety. A preceding vehicle distance measurement based on monocular vision for future intelligent traffic system is proposed. First, the model of internet of vehicles is proposed with the fusion of internet of things, intelligent recognition, and cloud computing technology. The vehicle can send back the location information and the preceding vehicle image to the internet of vehicles in real time, and request the nearby traffic signs and the geometry of the preceding vehicle. Then, establish a mathematical model of the monocular camera, and introduce a distance measurement based on monocular vision with a cooperation sign of traffic signs and lane lines. Finally, a preceding vehicle adaptive vision distance measurement is designed by comprehensive application of the distance measurement based on monocular vision. The simulation demonstrates that the distance measurement based on monocular vision is valid and effective, enriching driving assistance system.