Research Of BOPET Film Impurity Line Detector

With the image acquisition and digital image processing technology to continuously improve the level of research,film impurity detection devices constantly updated,the accuracy and high reliability BOPET film impurity line detector is not only demanded but also higher equipment requirements.In this paper,the algorithms based on LVQ neural network and hardware platform with TI's TMS320DM6446 processor chip embedded are used to the CCD image preprocessing of BOPET film to get the input feature values LVQ neural network,and by studying samples LVQ neural network learns with the ability to determine the target to enable the system to use for product quality testing and other areas in industrial production,in order to achieve automated production management.This paper firstly introduces the background and status quo of film impurity identification,and secondly the current film impurity detection technology and implementation methods were compared,analyzed this system to achieve the function and finalized based on the film impurity detection and recognition algorithm LVQ neural network online testing system solutions.In the system,select the ARM side of the Linux operating system and DSP end DSP / BIOS software development platform,using LVQ neural network algorithm film impurities target detection and identification.Hardware of the system is mainly part of the image acquisition module,data transmission module section,and embedded processing module score,external expansion memory module section,real-time data cache module section,in addition to the power module and the clock module section and other portions.This paper illustrates the porting and drive system,and the embedded software platform designed system software.Experiments show that LVQ neural network algorithm film impurity line detection system enables the film impurity detection and recognition,and is suitable for industrial field production line assembly line.