机器人CO2焊接熔池视觉传感与质量控制研究

Abstract: In the process of CO2 robot welding, the disturbances of strong arc light, noisy surface of work piece, variable pool shape and so on, makes it vary difficult to detect the pool shape with ordinary visual sensor, although these sensors have been successfully used in the process of GTA welding. In our research project, a special synchronous logic control circuit has been designed to control the start and close of CCD camera. Then, a novel welding pool image sensing system based on general industrial CCD camera has been developed. With this system, clear and amplificatory pool image has been acquired. Through the analysis of weld pool images, both size parameters and contour parameters, which could indicate the shape features of weld pool, had been defined to express weld pool information quantitatively. The relationships between weld pool shape parameters and welding process parameters have been investigated in details. All the progressed may be beneficial to the quality control of CO2 robot welding. Moreover, the achievements of visual image sensing technology and signal processing method under complex conditions in our project, have been successfully applied in the further researches of the aerospace science project, the National 863 Project and the Key Laboratory Foundation Project of National Defense Science and Technology.

研究通过工艺摄像机获取ＣＯ２焊接过程中熔池形状的图象信息，从中提取有关焊缝成形和熔透的规律，从而达到实时反馈质量控制的目标。拟采用先进的模式识别、信号处理、人工智能等相关技术，解决在熔滴过渡随机发生、电弧光闪烁剧烈、飞溅烟尘严重的恶劣环境下，分离目标背景，滤除干扰噪声，满足过程控制的速度和精度等关键问题，应用于实际生产。