基于机器视觉的车道线智能识别系统开发与实验研究

Development and Experimental Research of of Intelligent Lane Recognition System Based on Machine Vision

车道线智能识别系统的开发,能够为汽车自动驾驶和智慧城市建设提供必要的技术支持。传统车道线识别通过特征匹配算法和模型算法实现,具有较大的局限性。为此,笔者结合两者的优势,提出新型车道线智能识别系统,能够有效检测消失点并建立梯形DROI。将开发完成的智能车道线识别系统进行实践检验,识别成功率达到93.95%。由此可见,该系统基本能够满足使用需求。

The development of lane intelligent recognition system can provide necessary technical support for automobile driving and intelligent city construction. Traditional lane recognition is realized by feature matching algorithm and model algorithm, which has great limitations. For this reason, the author combines the advantages of the two, proposes a new lane intelligent recognition system, which can effectively detect vanishing points and establish trapezoidal DROI. The intelligent lane recognition system developed will be tested in practice, and the recognition success rate is 93.95%. Thus, the system can basically meet the needs of use.