基于机器视觉的智能贴附系统研究及其应用

Research and Application of Intelligent Adhesive System Based on Machine Vision

摘要: 在传统的电脑机箱面板加工过程中，由于面板上料在流水线上位置随意，几乎都是人工对面板进行贴附支撑垫和泡棉。针对这种方法生产成本较高，效率低下，无法满足高效生产需求的问题，提出一种基于机器视觉的引导定位技术，将视觉与工业机器人相结合，搭建一套智能贴附系统。该系统主要通过TCP/IP协议实现与机器人的通信，设计相机标定算法完成坐标转换，通过改进的模板匹配算法对面板特征识别、泡棉信息定位，利用卡尺工具算法检测贴附位置精准度，经过实地测试与验证，该系统表现出良好的运行稳定性、实时性和准确性，检测识别率高达99%，坐标定位误差在0.3 mm左右，最大限度地提高了生产效率。

Abstract: In the traditional process of processing computer case panels, due to the random position of the panels on the production line, almost all the support pads and foam are attached to the panels manually. To address the problems of high production costs, low efficiency, and inability to meet the demands of high-efficiency production, a guiding and positioning technology based on machine vision is proposed. This technology combines vision with industrial robots to build an intelligent attachment system. The system mainly communicates with the robot through the TCP/IP protocol, designs a camera calibration algorithm to complete coordinate transformation, uses an improved template matching algorithm to identify panel features and locate foam information, and employs a caliper tool algorithm to detect the accuracy of the attachment position. After on-site testing and verification, the system demonstrates excellent operational stability, real-time performance, and accuracy, with a detection recognition rate as high as 99% and a coordinate positioning error of approximately 0.3 mm, significantly enhancing production efficiency.