基于深度学习的VC散热片铜管焊接控制系统

Welding Control System of VC Radiator Copper Pipe Based on Deep Learning

针对人工控制VC散热片铜管焊接存在缺陷率高的问题,提出基于深度学习的VC散热片铜管焊接控制系统。该系统选用工业相机VCXU-53C采集焊接位置图像;利用改进的深度学习SqueezeNet轻量化神经网络建立焊接网络模型,实时分析采集的焊接位置图像,输出当前焊接进度;识别到焊接进度完成时,立即向高周波焊接设备发送停止信号,从而实现高周波焊接设备的自动化与智能化。该系统经实际应用验证:焊接良率可达到99.91%,满足实时检测的需求。

In view of the high defect rate of the manual control of the welding of the VC heat sink copper tube,a deep learning based VC heat sink copper tube welding control system is proposed.In this system,VCXU-53C industrial camera is used to collect the welding image;the improved SqueezeNet lightweight neural network is used to build the welding network model,analyze the collected image in real time and output the current welding progress.When the welding progress is recognized,the stop signal is sent to the high-frequency welding equipment immediately,thus realizing the automation and intelligence of the high-frequency welding equipment.The system has been verified by practical application,the welding yield can reach 99.91%,and it can meet the needs of real-time detection.