基于图像识别的移动机器人自动完成充电  

Automatic charging of mobile robot based on image recognition

作　　者：王建元[1] 郭威[1] 刘洋[1] 李菁华[1] 

机构地区：[1]东北电力大学电气工程学院,吉林吉林132012

出　　处：《电测与仪表》2017年第10期103-107,共5页Electrical Measurement & Instrumentation

摘　　要：移动机器人正在慢慢普及到我们的生活当中,而移动机器人在移动过程中消耗电能,当电能耗尽时需要人工为移动机器人进行充电,提出一种基于图像识别的方法,让机器人自动完成充电。移动机器人通过对图像的纹理分析和对兴趣点的快速匹配SIFT算法完成移动机器人识别电源插座,实现自动充电。本文的方法充分考虑到由于移动机器人的灵活多变性对所采集到的图像造成的影响,如图像旋转,图像周围有遮挡物体等,用纹理分析与SIFT算法识别插座,当所采集的图像与机器人本身记忆的图像达到一定匹配度时,移动机器人将完成自动充电。Mobile robot is slowly ＇spreading to our life, and mobile robot consumes the electric energy in the process of moving. It needs manual to charge for the mobile robot when electric energy is in depletion, so a method based on image recognition is put forward to let robot achieve the automatic charging. Mobile robot completes to identify power socket and realizes automatic charging through image texture analysis and point of interest fast matching SIFT algorithm. This method fully takes into account the influence such as image rotation, image around the occluded objects due to the flexible variability of mobile robot on the collection of image. The texture analysis and SIFT algorithm are used to identify the socket, and the mobile robot will complete automatic charging when the acquisition of the image and the memory image of robot itself achieve the matching degree.

关 键 词：移动机器人 纹理分析 兴趣点匹配 SIFT算法 自动充电 

分 类 号：TM93[电气工程—电力电子与电力传动]