Application of Automatic Recognition and Positioning Technology of Industrial Robot based on Monocular Vision Technology in Vehicle Manufacturing
参考中译:基于单目视觉技术的工业机器人自动识别定位技术在汽车制造中的应用


          

刊名：International Journal of Vehicle Structures and Systems
作者：Hongyu Chen（College of Mechanical and Electrical. Engineering., Shandong Vocational. College of Industry）
刊号：873HA008/I
ISSN：0975-3060
出版年：2023
年卷期：2023, vol.15, no.3
页码：446-453
总页数：8
分类号：U46
关键词：Monocular vision；Automatic identification；Positioning technology；Machine vision
参考中译：单目视觉;自动识别;定位技术;机器视觉
语种：eng
文摘：Combining machine vision technology with intelligent algorithms to improve the automatic identification and positioning technology of industrial robots to meet the production needs in different industrial environments is the current research direction of intelligent robots. In this study, an automatic identification and positioning technology for industrial robots based on monocular vision technology is proposed. First, the template matching algorithm and the Scale-invariant feature transform (SIFT) algorithm are introduced. Aiming at the shortcomings of the Hessian matrix in the SIFT algorithm in the process of eliminating boundary effects, an improved SIFT algorithm using Harris corner point detection is further proposed, and the improved SIFT algorithm is used in the automatic recognition and target positioning operations of industrial robots. In order to verify the performance of the proposed improved SIFT algorithm, the recognition accuracy and positioning angle deflection of the algorithm in different plastic sheets were detected. The experimental results show that under the improved SIFT algorithm, the recognition accuracy of five different styles of plastic sheets is above 98%. The improved SIFT algorithm also has less error between the predicted value and the actual value of the positioning angle deflection on the four plastic sheets. The robot under the improved algorithm is applied to the vehicle manufacturing industry and the production efficiency of the vehicle is improved through the automatic recognition and positioning technology of intelligent robot.
参考中译：将机器视觉技术与智能算法相结合，提高工业机器人的自动识别和定位技术，以满足不同工业环境下的生产需求，是当前智能机器人的研究方向。本文提出了一种基于单目视觉技术的工业机器人自动识别与定位技术。首先介绍了模板匹配算法和尺度不变特征变换(SIFT)算法。针对SIFT算法中海森矩阵在消除边界影响的过程中存在的不足，进一步提出了一种基于Harris角点检测的改进SIFT算法，并将其应用于工业机器人的自动识别和目标定位操作中。为了验证改进的SIFT算法的性能，测试了该算法在不同塑料板材上的识别精度和定位角度偏差。实验结果表明，在改进的SIFT算法下，对5种不同风格的塑料板材的识别准确率均在98%以上。改进的SIFT算法对四种塑料板材的定位角挠度的预测值与实际值之间的误差也较小。将改进算法下的机器人应用于汽车制造行业，通过智能机器人的自动识别和定位技术，提高了车辆的生产效率。



国家科技图书文献中心  全球文献资源网  京ICP备05055788号-26  机械工业信息研究院 2018-2024