Research On Fast Processing Method Of Glass Defect Image Based On Hadoop

The increase in the speed of the glass production line has led to an increase in the glass production per unit time,resulting in a large amount of data to be detected.In order to quickly and accurately process these large amounts of data,a glass defect detection method based on Hadoop is adopted in this paper.In order to realize efficient and accurate defect detection on the mass of glass produced on the belt,and since the Hadoop platform has high scalability,the glass defect detection method designed in this paper can be based on the production line when the speed or production method of the glass production line changes.The changes will quickly expand the existing architecture to match the new glass production line.The main tasks are as follows:(1)Based on the existing basic theoretical research of glass defect detection technology,the focus of the inspection of different types of glass defects and the basis of the system to identify the type of defects.(2)According to the glass defect detection method,a glass defect inspection program based on Hadoop platform is designed,which includes the software architecture cluster size design of image acquisition module and defect image module.(3)In the design of the image processing module,the main components and execution flow of the HDFS architecture and the MapReduce architecture under the Hadoop platform are mainly studied.The entire system architecture of Hadoop was optimized according to the glass defect detection method.(4)For glass defect detection using Hadoop architecture,the original defect location cannot be accurately recovered.A new glass defect detection method was designed using the SQL Server database,MD5 values of images,and Hadoop architecture.(5)Set up multiple types of Hadoop architectures,including Hadoop architecture with different cluster sizes in homogeneous mode and Hadoop architecture with different cluster sizes in heterogeneous mode.Several experiments were conducted and experimental data were collected.The experimental results show that the new detection method has higher detection speed than the traditional detection method,and can also restore the original position of the defect.