基于椭圆检测的充电口识别-《计算机技术与发展》

文章信息/Info

Author(s):: XIANG Bo-liang; YU Su; Shanghai University of Engineering and Technology,Shanghai 201620,China

Keywords:: ellipse detection; charging port identification; edge detection; arc segment screening; ellipse fitting

摘要:: 针对于新能源汽车的全自动充电技术,为了实现充电设备快速插接充电口的目的,提出了一种基于椭圆检测的充电口识别方法。该方法利用图像的预处理加强边缘,并结合 Canny 边缘检测得到边缘,将检测的边缘利用八邻域边缘跟踪筛选短弧, 以边缘梯度的正负性以及边缘的凹凸性为基准将检测边缘分为 4 类椭圆弧边缘。在此基础上,根据一定的限制条件以及椭圆的参数方程,选取满足限制条件的三段椭圆弧进行拟合,得到一系列的备选椭圆,通过特征参数之间的关系以及目标椭圆之间的联系,筛选出所需要的能够识别充电口的椭圆。在基于 Intel-Core 的 i7-6700HQ 处理器以及 Nvidia GTX 960M 的显卡的计算机设备上,通过 VS2017 的实验平台在多个角度进行多次实验,均能得到充电口的目标椭圆。表明该方法能够对插接的充电口做出准确并且快速的识别,为新能源汽车的全自动充电提供很好的先决条件。

Abstract:: Aiming at the automatic charging technology of new energy vehicles,in order to realize the quick plugging of charging port for the charging device,a charging port identification method based on ellipse detection is proposed. The method strengthens the edge with pre-processing of the image and obtains the edge by combining the Canny edge detection. The detected edges are screened by the eight-neighbor edge tracking to select the short arc. With the positive and negative of edge gradient and the convexity of the edge as the reference,the detection edges are divided into four types of elliptic edges. On this basis, according to certain constraints and the parametric equation of the ellipse, a three-segment elliptical arc satisfying the constraint condition is selected to obtain a series of candidate ellipse. Through the relation between characteristic parameters and the relation between target ellipse, the ellipse which can recognize the charging port is screened out. On the computer based on Intel-Core-i7-6700HQ processor and the Nvidia GTX 960M graphics card,the multiple experiments can be performed in the VS2017 experimental platform at multiple angles,and the target ellipse of the charging port can be obtained. It is indicated that the method can provide an accurate and fast identification on the plug-in charging port,which provides a prerequisite for fully automatic charging of new energy vehicles.