单目视觉智能车路径识别及控制策略研究

Track Identification and Control Strategy of Smart Car of Monocular Vision

研究了基于CMOS摄像头的图像采集方法,以及智能车赛道路径识别。提出了自适应差分边缘检测算法,采用取点求面积的方法提取指引线的相关参数。自适应差分边缘检测算法是在一般的边缘检测算法的基础上提出的,它能根据提取的左右边缘存在情况调整搜索范围、阈值,以及差值的求取方法。使用海伦公式求指引线上所取的三角形的面积,据此提出了1种基于三角形面积的智能车速度控制方法,此方法以指引线上的三角形面积反映赛道的弯曲程度,并以此作为智能车速度控制的控制变量。

Abstract：The method of image acquisition and track identification are studied. A kind of adaptive differential algo- rithm for edge detection is put forward. The algorithm is proposed on common differential algorithm for edge detection, which can adjust the range of search, threshold, and the method of calculating difference, Thus, this algorithm is provid- ed with more anti jamming capability. Heroffs formula is used to take the area of the triangle on lines, arid a simple can-trol method of speed of smart car is used. The area of triangle on curve reflects the bend of track, which is taken as the control variable of speed control of smart car.