Development Of Online Detection System Of Inner Wall Defects Of Gas Cylinders Based On Machine Vision

Aluminum alloy cylinders are the main body of aluminum alloy seamless gas cylinders.After being processed into aluminum alloy seamless cylinders,aluminum alloy cylinders can be used to refill various kinds of rare gases,high purity gases,standard gases and special gases in various fields,which belong to industrial gas cylinders.Due to the direct contact between the inner wall and the gas,if the inner wall defect is large,it may cause serious corrosion and other effects,resulting in greater economic losses.At present,the detection of the inner wall defects of aluminum alloy cylinders is still manual detection at home and abroad.The manual detection is not only inefficient,but also because of the long fatigue and psychological factors of human eyes,missed detection and leak detection are difficult to avoid,and it is difficult to meet the requirements of quality inspection.Therefore,the quality detection system of the inner wall of the cylinder based on machine vision is of great significance.To solve the problems above,a complete on-line inspection system for cylinder wall defects based on machine vision is proposed in this paper.The contents of this paper are mainly focused on the following aspects:The main defects of the inner wall of the aluminum alloy cylinder are briefly explained,and the characteristics of the main defects are analyzed and summarized.The key and difficult points in the on-line inspection system of the inner wall defect of the cylinder based on machine vision are summarized.In the hardware system,the imaging system is designed mainly,and the load system is briefly introduced.To choose appropriate light source,camera,lens for the imaging system,and the design of the supporting structure of the imaging system,which is convenient for fixing camera and implementation in the subsequent motion program support structure,drive the camera,lens and light source movement;loading systems for supporting cylinders,and realize rotation rules in the control program under the subsequent movement.In the software system,mainly to write the MFC interface,to display the defect images,and other major regulatory factors and data in detection system;through the interface control,set and adjust relevant parameters,to achieve a stable uniform cylinder intermittent rotation of the loading mechanical platform system,to uniform advance and retreat steadily of the imaging system,the two jointly coordinated action to achieve real-time image acquisition,follow-up by calling the algorithm,can realize the online image processing,detection and judgment of defects.In the algorithm,using theimage processing method based on Canny operator,through image processing software Halcon to complete the image processing,introduces the detection method of aluminum alloy pits and cylinder wall in transverse and longitudinal scratches strain defects.Canny operator is used to select two kinds of edge detection for different parameters according to this three kinds of defects specific features to extract the edge part contains the main defects.Then connected,screening and merging operations according to the basic characteristics,determine whether to comply with the final recognition type gas wall three defect standard.The proposed detection method has been tested,and the average time of detecting a picture is less than one second.When detecting a gas cylinder,the total time is not equal according to the diameter and length of the inner wall of the cylinder.The detection time meets the requirements of the production enterprise and has low missed detection and leak detection rate.It can realize visual online detection,which has certain practical application value.