港口机械用攀爬机器人视觉导航定位方法研究

为研究攀爬机器人在港口机械金属结构表面的自主移动,提出了一种应用于攀爬机器人的视觉导航定位方法。针对港口机械金属结构复杂,表面存在油污和阴影,解决现有的视觉边缘识别方法对金属结构边缘识别效率低且不连续的问题。首先通过自适应阈值的Canny进行边缘特征提取;之后根据边缘线角度特征建立约束关系,利用Hough变换剔除干扰线,并用最小二乘法拟合最终金属结构边缘;求取结构中心线方程和运动方程并计算偏航角和偏距,实现对机器人的运动控制。以港口机械某一支腿为例进行试验,实验数据表明,该方法可以很好地引导机器人的运动,满足实际应用。

Visual navigation and positioning method of climbing robot for port machinery

In order to study the autonomous movement of climbing robots on the metal structure surface of port machinery,a visual navigation and positioning method for climbing robot is proposed.Considering that the metal structure of port machinery is complex,and there are oil stains and shadows on the surface,in order to solve the problem of low efficiency and discontinuity of metal structure edge recognition in existing visual edge recognition methods,firstly,edge features were extracted by Canny algorithm with adaptive thresholds;then,according to angle characteristics of edge lines,the constraint relationship was established,the interference line was removed through Hough transform,and the final metal structure edge was fitted by least square method;the centerline equation and motion equation of the structure were obtained,and the yaw angle and offset were calculated,so as to control the motion of the robot.Taking a supporting leg of a port machinery as an example,tests were conducted,and test data shows that this method can guide the robot movement well and meet relevant practical application requirements.