分布式算法对工业机器人视觉伺服图像处理系统的影响  

Influence of Distributed Algorithm on Visual Servo Image Processing System of Industrial Robot

作　　者：方赟 李石朋 王娟[3] Fang Yun;Li Shipeng;Wang Juan(China International Marine Containers(Group)Ltd.,Shenzhen,Guangdong 518067,China;School of Mechanical Engineering Zhejiang University,Hangzhou 310058,China;School of Automation Science and Engineering,South China University of Technology,Guangzhou 510641,China)

机构地区：[1]中国国际海运集装箱(集团)股份有限公司,广东深圳518067 [2]浙江大学机械工程学院,杭州310058 [3]华南理工大学自动化科学与工程学院,广州510641

出　　处：《机电工程技术》2022年第12期122-126,225,共6页Mechanical & Electrical Engineering Technology

摘　　要：通过视觉传感器使机器人与外部环境进行交互的机器人视觉伺服技术,为机器人的自主感知、自主决策等智能化功能的构建提供了方案。在传统以上位机为核心的集中式数据处理架构中,传感器带来快速与高质量图像信息的同时,海量的图像数据也对上位机处理数据的实时性和可靠性提出了挑战。综合考虑机器人视觉伺服图像的数据处理传输时易发生拥堵、丢包等缺陷和雾计算分布式技术的实时优势,提出了一种基于雾计算的视觉伺服系统图像实时自适应处理架构,利用传感器将获取到的图像数据传输到处理节点,根据不同频率和大小的图像数据算法进行计算资源的自适应分配,在图像处理单元得到完整的图像数据的情况下保证完成图像的实时处理。同时利用实验室视觉伺服平台和雾计算节点设计了相关实验方案,并进行了算法对比验证,结果显示在其他条件相同的情况下,文中算法较传统算法在图像处理效率方面至少能够提高50%,表明所提算法架构的有效性。The robot visual servo technology,which enables the robot to interact with the external environment through visual sensors,a solution was provided for the construction of intelligent functions such as autonomous perception and autonomous decision-making of robots.In traditional centralized data processing architecture with the upper computer as the core,while sensor brings fast and high-quality image information,massive image data also challenges the real-time and reliability of data processed by upper computer.the defects were comprehensively considered such as congestion and packet loss in the data processing,transmission of robot visual servo images and the realtime advantages of fog computing distributed technology,a real-time adaptive processing architecture was proposed for visual servo system images based on fog computing.The acquired image data was transmitted to the processing node,and adaptive allocation of computing resources was performed according to the image data algorithm of different frequencies and sizes,then the real-time image processing was ensured when the image processing unit obtains the complete image data.At the same time,the experimental scheme was designed,also the algorithm was compared and verified by using the laboratory visual servo platform and fog computing nodes.The results show that under the same conditions,the algorithm in this paper can improve the image processing efficiency by at least 50%,illustrates the effectiveness of algorithm architecture proposed in this paper.

关 键 词：分布式计算 工业机器人 图像处理 自适应算法 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]