列车风管管接头间接识别定位方法

Indirect Identification and Location Method of Train Air Duct Joint

为完成移动机器人对列车风管的摘解任务,需要对列车风管管接头进行定位。针对列车风管管接头不易识别且识别定位需要一定实时性和精度的问题,利用双目视觉系统提出了一种间接识别定位方法。在目标识别中,根据列车风管的特点,基于颜色特征在HSV空间中对风管近端金属圈进行识别,提取出金属圈形心坐标;在三维重建中,利用具备较高精度和一定实时性的组合立体匹配算法对风管近端金属圈进行三维重建,得到其三维坐标后根据管接头与近端金属圈相对位置关系固定的特点,通过空间坐标转换得到管接头的三维坐标,实现列车风管管接头的定位。最后,在(200~650)mm的距离范围内进行实验验证,实验最大绝对误差为12.88mm,算法最长耗时为245ms,结果表明该方法满足要求。

In order to complete the task of extracting the air duct of the train by the mobile robot,it is necessary to locate the joint of the air duct of the train.Aiming at the problem that the joints of train ducts are difficult to identify and the identification and positioning require a certain degree of real-time and accuracy,an indirect identification and positioning method is proposed based on the binocular vision system.In target recognition,according to the characteristics of the train duct,the metal circle at the proximal end of the duct is identified in the HSV space based on the color feature,and the centroid coordinates of the metal circle are extracted;in the 3D reconstruction,the use of high precision and certain real-time The combinatorial stereo matching algorithm performs three-dimensional reconstruction on the proximal metal circle of the duct.After obtaining its three-dimensional coordinates,according to the characteristics of the fixed positional relationship between the pipe joint and the proxi-mal metal circle,the three-dimensional coordinates of the pipe joint are obtained through spatial coordinate conversion to realize the train Positioning of duct joints.Finally,the experiment was verified in the distance range of(200~650)mm.The maximum absolute error of the experiment was 12.88mm,and the longest time-consuming algorithm was 245ms.The results showed that the method met the requirements.