Automatic Detection Of Silver Contacts Welding Quality Based On Machine Vision

The signal relay can utilize small electrical signals to control the relatively large object in the execution circuit,control multiple loops and objects,and achieve distance control.It has good switching performance,little closing impedance and big disconnecting impedance.With the failure-to-safety guiding performance,it can control multiple loops and have strong anti-lightning performance,little noise and small temperature response.It is largely used in the system configured with relay technology,and it is also used as the interface unit in the system consisting of electronic components and a microcomputer to connect the host of the system with execution units such as model machine,track circuit and switch.Because the quality of welding of the relay's silver contacts is the main factor causing the failure,the welding strength of silver contacts in the manufacturing process of the relay has to pass strict testing.At present,Chinese manufacturers mainly use shear force tools for sample testing of each batch of silver contacts to assess their quality.This method not only has low efficiency,high labor intensity and low resistance to subjective factors but also cannot achieve the quality testing of each product and has difficulties in meeting the requirements of modern flow-line production of the relay.This paper analyzed the research background of the welding quality of silver contacts of the signal relay as well as related applications and researches of machine vision technology in the field of industrial inspection.In consideration of current relatively laggard automation degree of welding quality test of silver contacts,this paper proposed a method of welding quality test of silver contacts based on machine vision technology.Considering the actual situation of the production line,the visual inspection system was introduced to the station to achieve real-time online testing of welding quality of silver contacts as well as test of errors,repeatability and stability of the inspection system.The main work included:On the basis of the demand analysis,the overall plan based on the embedded visual inspection system was proposed,and according to the testing requirements,the selection of hardware of visual inspection system,including camera,lens and light source as well as the erection of the platform was completed.In the aspect of image processing algorithm design,because silver contacts had low image contrast and noise,image enhancement and image filtering algorithms were introduced,and after analysis and comparison,the gray stretch image enhancement algorithm and the Gaussian filtering algorithm were selected.For image feature location,the PatMax template localization algorithm based on geometric shapes was selected.Depending on the specific inspection content,threshold segmentation and edge detection were respectively carried out for the image.The programming of the image processing algorithm of silver contacts was completed at the software platform of In-Sight.Stability and inspection accuracy of the visual system were verified by the standard part produced.The experimental result showed that the use of machine vision technology could accurately and quickly complete the inspection of the welding quality of silver contacts and had important significance to improve the automation degree of production.