Research On Visual Inspection Technology Of Valve Plate Surface Defect

The valve plate is an important basic part of the air conditioning compressor,and its surface defects will directly affect the normal operation and service life of the whole system.It is necessary to detect the valve plate defects before leaving the factory.At present,the detection of valve plate surface defects mainly stays in the stage of manual detection.Manual inspection is not only time-consuming and labor-consuming,but also difficult to detect many kinds of subtle defects on the valve plate by naked eyes.With the development of automation technology and the actual needs of valve plate manufacturers,it is of great practical significance to design and manufacture an automatic inspection system for valve plate defects.The surface of the valve plate has a grinding grain with the same direction and different depth.The reflection of knife pattern on different angle light source will seriously affect the image quality.At present,machine vision has been widely used in the detection of surface defects of parts,but there is no perfect detection method for the valve plate,which has a serious impact on the image quality.In order to overcome the influence of knife reflection on the detection process,an automatic detection system for valve plate defects is designed and manufactured.The following work is done:1.According to the characteristics and distribution of valve plate surface defects,the defects were divided into five categories: Marginal abrasive,Scratch,Press hole,Pockmark and Defect.According to the accuracy requirements of quality inspection standards,chose appropriate hardware equipment such as light source,camera,lens,computer,etc.2.Through the observation of a large number of valve plate images collected under various lighting modes,it was found that the overall area of valve plate is clear and the outline is obvious under the plane light,but some minor defects are difficult to be found,and it is easy to cause omission only by the plane light source;Low angle parallel light source(hereinafter referred to as "directional light")can effectively highlight the characteristics of minor defects.However,when the directional light source is at a special angle with the Grinding grain,the reflection of the Grinding grain will cause the image tobe overexposed or too dark.In view of the above situation,a lighting method combining plane light and directional light is designed to take advantage of the advantages of the two light sources.3.A set of hardware system was designed according to the position of light source and camera,which realized automatic lighting,image acquisition and preservation,defect extraction,motion control and other functions.Aiming at the reflection of Grinding grain,the reflection law was found,and the average gray level of valve plate area was ensured to be in the proper range by adjusting the exposure time of camera in real time.4.According to the distribution characteristics of defects and the reflective characteristics of different lighting methods,a detection method combining plane light detection and directional light detection was designed.Aiming at the bright spots that can not be eliminated in the image,a method of finding defects based on the edge line was proposed: according to the position and direction of the gray abnormal area and edge line,judge whether the abnormality was caused by the defects or by the bright spots,effectively avoiding the false detection.5.A set of program for the detection of valve plate surface defects was designed and developed.The operator can control the detection system through the graphical interface of the software to carry out image acquisition,light switching,defect detection,parameter setting,image storage,etc.6.Designed the experimental control group,developed the test standard and pass the experimental test system.The experimental results were compared with the test standards,the test time was recorded,and the performance of the system was evaluated.The results show that the detection system has good stability,detection speed is slightly lower than that of human,and detection accuracy is high,which is in line with the actual production needs.At the same time,the detection algorithm used in this system provides a new idea for the surface defect detection of valve plate,such as stamping and grinding plate parts.