Transparent Glass Panel Defection Of Image Acquisition Software System Research

In industrial production,intelligent mobile screen cutting and re-process will inevitably lead to quality problems such as scratches,chipping,and other foreign matter.These problems not only affect the consumers of the product,but also affect the consumer's choice of the brand recognition.So,how to screen out defective products from a large number of glass panels is particularly important.The traditional artificial has low detection efficiency,high labor intensity,the test results can not be digitized and met the business requirements of the rapid production.To this end,from batches important issue transparent glass plate production quickly and accurately defects acquired in a product to be solved.In order to realize the fast,accurate and efficient on-line inspection of glass panel surface,the software system of laser scattering image acquisition based on FPGA is designed by laser scattering imaging principle.The main work is as follows:(1)The architecture of the C/S system is presented,which takes the server as the control host and the image data acquisition card as the image acquisition and processing unit.(2)Using the high parallel characteristic of FPGA,high transmission speed and fast response characteristic of GigE Vision network,the design and function of image acquisition system are completed.(3)According to the characteristics and types of glass panel surface defects,a set of image processing algorithm based on artificial neural network and suitable for glass panel defects is designed and researched,which has the advantages of good adaptability and high classification efficiency.(4)The verification platform is built to further verify the system.Experimental results show that the software can image well under the condition of 30mm/s,the communication speed of Ethernet is 5.3Mbytes/s,and the response time of the card to the server is about 0.3ms.The average false detection rate is about 5.09%,and the average undetected rate is about 1.05%.The glass panel image acquisition software system designed in this paper has the characteristics of good response characteristics,high real-time performance,high degree of automation and strong data throughput,which meets the requirements of glass panel surface defect acquisition system.It provides a reference for the popularization and development of machine vision technology in industry.