| Machine vision is a automatic and intellectual technology which capture images in realistic scene by image sensing device and then identify and storage the features in the picture using computer,sometimes generating the corresponding feedback signal to act on the measured object.This technology has the advantages of high processing speed and high recognition accuracy.It can detect small objects which are difficult to be distinguished by human eyes.The defects produced by glass edge grinding are very slight,it is easy to produce false detection or miss detection relying on the human eyes.In view of this situation,this paper applies machine vision technology to glass edge defect detection to develop a stable,reliable and efficient defect detection method.This research is carried out from three angles: theory,practice and result verification.The theoretical aspects includes the analysis of the present situation of machine vision at home and abroad,the research of light source,the process of glass edge grinding,the forming mechanism of glass defect,the types of defects(including bright spots,explosive edges and white lines),imaging principle of Industrial CCD camera,and the image processing techniques(image denoising,image sharpening,image segmentation,edge detection,mathematical morphology,and image file storage format);In practice,the development of automatic detection system for glass edge-grinding defects is carried out.Including the formulation of the overall plan and the detailed design of the system.The content of the overall scheme includes the functional modularization of the system,the overall architecture diagram of the system,the selection of the sensing device structure,the design of the transmission link and the configuration of the calculation module.Detailed design includes selection of hardware,design of defect recognition algorithm and software,and analysis and processing of subsequent data.The algorithm,based on the existing image processing algorithm,aims to realize feature extraction of bright spot,crack edge and white line.Object-Oriented programming language used in Software Design,it improves the coding reusability,expansibility and maintainability of image processing algorithm program.The contents include the partition of system use cases,the design of data flow,the design of internal call structure,the design of message passing process and the design of system interface.At present,most of defect detection techniquesin machine vision do not store the detected data,which will lead to the waste of data,and it is not conducive to the statistical analysis and trend prediction of defects by decision makers.Therefore,an open database connectivity technique is introduced in this paper to preserve the characteristic data of defects.In the result verification aspect,comparing the data characteristics of each defect detected by the experimental system with the acceptance standard,the results show that the detected data are within the range of defect standard value,it is proved that the system is an effective technique for detecting the defects of glass edge grinding.The function evaluation of the system is carried out,the accuracy,the number statistics,the dimension calculation,the position calculation,the mixed detection and so on are tested respectively,the application value of the system is proved.Finally,the performance of the system is evaluated.The system’s defect detection speed,defect location recording ability,data storage capacity and statistical ability of data analysis are evaluated.The efficiency,reliability and stability of the system are proved.In view of the research of the machine vision defect detection,the main method used is the combination of theory and practice.The first half pays attention to the accumulation of theoretical knowledge,and the second part devotes itself to the design and development of the system.The idea of development is the combination of software and hardware.It has successfully solved the present situation that the defects of glass edge grinding are difficult to detect.At the same time,it lays the foundation for the interconnection of all things in the Internet of things times. |
