摘要
电动汽车自动充电机器人是解决电动汽车充电难,充电操作不便的重要途径之一。设计了一种基于椭圆检测和双目视觉的电动汽车直流充电口位姿估计程序,该程序首先需要对图像进行预处理操作,然后进行轮廓提取与筛选,最后进行最小二乘法椭圆拟合、过滤,并针对拟合过滤后的椭圆进行电动汽车直流充电口位姿的估计。经过针对该方法程序的第三方测试,得出的结论是精确度满足需求,并且算法稳定,应用场景普适性高。
Electric vehicle automatic charging robot is one of the ways to solve the difficult and inconvenient of charging operation of electric vehicles.A pose estimation method of electric vehicle DC charging port was designed based on ellipse detection and binocular vision in this paper.Firstly,the image is preprocessed,then the edges are extracted and filtered,and finally the least square ellipse fitting is carried out after filtering,the pose of the DC charging port is estimated according to the ellipse after filtering.Through the third-party test,it could be found that the accuracy of the method meets the requirements,and the algorithm is robust and the application scenario has high universality.
作者
刘文奇
胡晓羽
Liu Wen-qi;Hu Xiao-yu(Center Academy of Shanghai Electric Group Co.Ltd.,Center Academe Shanghai 200040,China)
出处
《包钢科技》
2021年第5期91-95,共5页
Science & Technology of Baotou Steel
关键词
电动汽车
机器人导航
位姿估计
双目视觉
椭圆拟合
electric vehicle
robot navigation
pose estimation
binocular vision
ellipse fitting