视觉引导下物流模型车的边缘检测与路径识别  

Edge Detection and Path Recognition of AGV Model Car under Vision Guidance

作　　者：刘泽伟 冯靖靖 李婉钰 姜梅雪 吴俊华[2] LIU Zewei;FENG Jingjing;LI Wanyu;JIANG Meixue;WU Junhua(College of Engineering,Qufu Normal University,Rizhao Shandong 276826;College of Computer Science,Qufu Normal University,Rizhao Shandong 276826)

机构地区：[1]曲阜师范大学工学院,山东日照276826 [2]曲阜师范大学计算机学院,山东日照276826

出　　处：《软件》2022年第7期39-42,共4页Software

基　　金：国家大学生创新创业训练项目(202110446162)。

摘　　要：为解决物流领域中的自动引导车辆(简称AGV)在视觉引导方式下实时性差、识别精度低的问题,选用智能车作为AGV的模型车,来进行图像处理和路径识别的算法研究。为提高对道路边缘的快速识别,将原有的边缘检测法改进为一种多状态边缘检测算法,即考虑一般道路下的多类型图像状态。针对特殊路径的识别,使用陀螺仪角速度积分法来规避车速对标志位的影响,并通过对照实验证明了该算法的有效可行。经测试和应用,证明了一般情况下图像处理运算速度的提升和特殊路段识别的高精确性。In order to solve the problems of poor real-time performance and low recognition accuracy of automatic guided vehicles(AGVs)in the field of logistics in the visual guidance mode,intelligent vehicles are selected as AGV model vehicles for image processing and path recognition algorithm research.In order to improve the fast identification of the road edge,the original edge detection method is improved on a multi-state edge detection algorithm,that is,the multi-type image states under the general road are considered.For the identification of special paths,the gyroscope angular velocity integration method is used to avoid the influence of vehicle speed on the sign position,and the effectiveness of the algorithm is proved through control experiments.After testing and application,it is proved that the speed of image processing and the high accuracy of identifying special road sections are improved in general.

关 键 词：自动引导车辆 CMOS摄像头 动态阈值 边缘检测 角速度积分 对照实验 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置]