Research On Intelligent Detection Algorithm And Equipments Of Solder Joint Based On Deep Learning

In the production of welding parts,the number of solder joints is large,easy to leak welding,resulting in quality problems.At present,the detection of the number of solder joints is mainly manual,the labor intensity is large,and prone to error,so it needs an automatic and accurate detection of the number of solder joints device.This topic to detect the number of solder joints on the surface of welding parts as the main target,design and manufacture 1 can intelligently detect the number of solder joints on the surface of welding parts equipment.Through a large number of literature research and welding defect research,the mechanical structure of the inspection equipment and the overall framework of the inspection system are constructed.On the equipment characteristics of the color of the light source of the image acquisition system and lighting mode are analyzed,and determine to use blue light source and the way of positive direct lighting,then the camera lens and other optical parameters of imaging system,determine the detection under the condition of using 8 mm lens,a 200 mm distance,aperture of F8 is optimal acquisition scenarios.After the acquisition of image,the grayscale processing research,and study the image noise,filtering and edge detection of image processing algorithm,analyses the method of least square circle fitting and hoff round transformation of solder joint in this topic in the fitting test,and test,detection effect is very poor and the rate of false positives and non-response rates can fluctuate.Then the target detection method of deep learning neural network is studied,and the detection rate of false positives is only 1.8% and the missing alarm rate is 0%.Finally,based on the above detection methods,the visual detection equipment of welding joints is built,and the equipment of the whole machine is completed through image acquisition system,linear motion unit,control system,detection system,etc.,and it enters the production line of the factory for quality inspection.