Research On MGLR Control Chart With Image Calibration In Machine Vision System

Manufacturing is the basic industry of the national economy.The development level of the manufacturing industry reflects the productivity level and national's overall strength.With the development of science and technology,the manufacturing industry is developing towards intelligent manufacturing.Under this general background,how to improve product quality is important for the country to improve the core strength of the manufacturing industry.In the production process,people determine whether the quality of the product is qualified or not according to whether the deviation of the product's quality characteristic is within a reasonable range.Traditionally,products' surface quality is inspected manually.Manual inspection is affected by many factors and does not meet the requirements of intelligent manufacturing.Since the image information contains information about the products' surface quality,the control of the products' surface quality can be transformed into the control of the image quality.The machine vision system(MVS)can efficiently obtain images which contain quality characteristics such as product geometry,surface defects,surface finish,etc.,is becoming more and more widespread in intelligent manufacturing.Some products' surfaces,such as LCD screens,ceramic tiles,etc.,have the characteristic that the detection characteristics follow a certain distribution(such as trying to detect uneven defects in LCD monitors)or follow specific patterns(such as ceramic tiles).To detect whether the quality characteristics of the surface of these products are offset,this article uses a Multivariate General Likelihood Ratio control chart based on the image monitoring method to monitor the quality of the image.This method can detect the image with multiple faults,and can also estimate the scale,the size,and the change points of the faults.This important diagnostic information can help reducing downtime.The traditional MGLR control chart assumes that the difference between the images to be detected and the images in a known controlled state only comes from the quality difference,while external conditions do not affect the image.However,in the intelligent manufacturing production process,there will be the impact of the placement of the product to be tested and the change in the detection environment light,etc.,resulting in inaccurate detection results,which has a greater impact on product quality control.This paper presents the MGLR control chart with image calibration under the machine vision system,hoping to minimize the impact of the external environment on the product quality control process,expand the scope of use of the MGLR control chart in smart manufacturing,and to improve efficiency,reduce labor costs and total While maintaining the cost,maintain the accuracy of the MGLR control chart.This paper uses numerical experiments to study the performance of the MGLR control chart with image calibration under the machine vision system.Regardless of the number of offset centers,the MGLR control chart with image calibration under the machine vision system can detect defects more effectively,and in the case of a smaller degree of offset and/or offset range,the control chart has better performance.