基于立体视觉的安全车距新型识别技术的研究

Research On New Recognition Technology Of Safe Car Distance Based On Stereo Vision

车辆行驶距离是交通安全的重要指标之一。安全车距的测量已成为智能交通领域的一个日益热门的研究领域。通过分析立体视觉的基本原理,标定CCD传感器的内外参数,提出了一种测量立体视觉的方法,建立了基于立体视觉和DSP的车辆距离控制系统。一旦车辆的速度和距离形成非安全关联,它就会发出警告,并上传数据或强制限速。该方法根据从左右传感器获取的目标车辆图像的不同坐标,识别特征点,计算到目标车辆的距离,分析车辆距离的安全性。实验结果表明,该方法测量范围宽,测量精度高,运算速度快,能够满足智能交通中安全车距测量的实际需要。

Vehicle distance is one of the important indexes of traffic safety. The measurement of safe distance between vehicles has become an increasingly popular research field in the field of intelligent transportation. By analyzing the basic principle of stereo vision and calibrating the internal and external parameters of CCD sensor, a method of measuring stereo vision is proposed, and a vehicle distance control system based on stereo vision and DSP is established. Once the vehicle’s speed and distance form an unsafe relationship, it will issue a warning and upload data or impose a speed limit. According to the different coordinates of the target vehicle image obtained from the left and right sensors, the feature points are identified, the distance to the target vehicle is calculated, and the safety of the vehicle distance is analyzed. The experimental results show that the method has a wide measurement range, high measurement accuracy and fast calculation speed, and can meet the actual needs of the safe distance measurement in intelligent transportation.