Research And Implementation Of Injection Molding Bottle Defect Detection System Based On Machine Vision

With the continuous improvement of social productivity level,people pay more and more attention to medical health and food safety issues.As the mainstream of drug packaging,the quality of injection bottles plays a vital role in the transportation,storage,sale and use of drugs,directly affecting the safety and effectiveness of drugs.At present,the quality inspection of injection bottles in China mainly relies on online manual inspection and off-line sampling inspection,which is not only inefficient and costly,but also affected by the personal experience and working status of operators.The inspection method based on machine vision has incomparable advantages in speed,accuracy and objectivity.Therefore,this paper studies the inspection system of injection bottle based on machine vision technology,and realizes the defect detection of injection bottle through image processing algorithm.The main contents and innovations of this paper are as follows:(1)In this paper,the image preprocessing technology of bottle is studied in depth.Combining binarization algorithm and Freeman chain code,the contour of the bottle is extracted,and the image of the bottle is corrected by calculating the minimum outer rectangle of the contour and affine transformation.In the image enhancement part,Gamma transform is improved to make adaptive adjustment according to the overall brightness of the image.In the part of filtering and denoising,several main filtering and denoising algorithms are simulated and selected.(2)Aiming at the bottle image segmentation under non-uniform illumination,several classical global adaptive threshold segmentation algorithms and local adaptive threshold segmentation algorithms are studied.On this basis,combined with the gray-scale distribution characteristics of bottle defects,the threshold segmentation algorithm based on neighborhood mean is improved.The weighted mean of neighborhood is calculated quickly by integral graph to realize the bottle image segmentation.The simulation comparison verifies the superiority of the improved algorithm in segmentation speed and anti-noise performance.At the same time,a defect detection algorithm based on connected region labeling is designed for the segmented bottle image,and a defect detection algorithm based on sliding window is designed to detect the end face of bottle mouth and the contour of the outer bottle mouth such as screw teeth.(3)The overall architecture of the system is studied,and the design and implementation of the hardware and software parts of the system are completed respectively.The hardware part introduces the hardware composition of the whole system,studies and analyses the selection of relevant equipment of image acquisition module and execution control module,and the software part completes the design and implementation of the whole detection process,and completes the development of GUI.(4)Finally,the defect detection system is tested and analyzed from the two aspects of accuracy and speed of defect detection.The experimental results show that the system designed in this paper can complete the task of on-line detection of bottle defects.The accuracy rate is over 90%.The system runs steadily and has practical value.