基于场景语义信息分析的AGV视觉导航技术研究  

Research on AGV visual navigation technology based on scene semantic information analysis

作　　者：焦科杰 陈颖 丁尧[2] Jiao Kejie;Chen Ying;Ding Yao(Nanjing Panda Information Industry Co.,Ltd.,Nanjing 210000,China;Nanjing University Jingling College,Nanjing 210000,China)

机构地区：[1]南京熊猫信息产业有限公司,江苏南京210000 [2]南京大学金陵学院,江苏南京210000

出　　处：《江苏科技信息》2020年第9期67-70,共4页Jiangsu Science and Technology Information

摘　　要：为了提高AGV视觉导航的自主性,文章在传统的AGV视觉巡线导航技术中,通过增加对场景语义信息提取,提出了基于场景语义信息分析的导航路径选择方法,即在获取场景的图像信息后,提取场景中的导航线信息和场景中的语义信息,通过语义信息的分析与匹配,选择合适的路径,通过图像分割技术,保留需要的路径信息,实现对导航路径的选择,从而减少冗余信息对导航的影响。该方法提高了AGV视觉导航的自主性,在减少导航信息的基础上,也提高了AGV导航的鲁棒性。To improve the autonomy of AGV visual navigation,this paper puts forward a navigation path selection method based on scene semantic information analysis in traditional AGV visual patrol navigation technology,after obtaining the image information of the scene,the navigation line information and the semantic information in the scene are extracted,and the combination is selected through the analysis and matching of semantic information.The appropriate path,through image segmentation technology,preserves the required path information,and realizes the selection of navigation path,so as to reduce the impact of redundant information on navigation.This method improves the autonomy of AGV vision navigation,reduces the navigation information,and improves the robustness of AGV navigation.

关 键 词：自动导航车 视觉导航 场景信息 语义分析 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]