Research And Implementation Of Surface Defect Detection Of Wood Door Spray Paint Based On FPGA

The detection of defects in wooden door spray painting process is an important issue that needs to be solved in the wooden door painting process.At present,the domestic small and medium-sized enterprises mainly adopt the artificial way to carry out the Lacquer defect detection of wooden door.This method of defect detection has many defects like inefficient,low degree of automation and low accuracy.In response to this problem,this paper developed a Lacquer defect detection system of wooden door based on FPGA.The main research work of this paper is as follows:According to the spray painting process,the wooden door surface defects were detection.first,the image of the wooden door surface were collected and Pre-processed.then,the traditional Sobel algorithm was used for image edge detection,to acquire the edge image that meet the requirements.In this paper,the median filter is used to replace the average filter to get an improved Sobel algorithm and a more accurate edge detection image.On the basis of this,the corrosion expansion algorithm is analyzed.The Matlab simulation verifies that the image edge detection meets the requirements.The defect of painted wooden door surface was extracted.Based on the experimental data,we found that the aspect ratio features and complexity features can effectively distinguish the types of target defect features.Therefore,by using them as the input eigenvectors of the classifier,a BP neural network classifier suitable for the research subject is designed to classify the surface defects of wooden door which include the defects of scar and crack.This article selected Cyclone ? series EP2C8Q208C8 which manufactured by Altera company as the core control system to achieve the wooden door surface image data collection and processing.Hardware circuit mainly include the image acquisition circuit,FPGA core circuits,image processing circuits,and display circuits.The camera was used for data collection of detected wood door surface,and then transmitted to the FPGA to process it.the whole system was completed which include image data collection and processing of the wooden door surface,the recognition of image edge and the wooden door defects,and VGA Monitor for terminal display.The design of software part was based on SOPC technology,by using Quartus ?software the top file was created,to achieve SOPC system hardware Design,the configuration of Nios soft processor and customization of corresponding IP cores such asJTAG,EPCS,SDRAM,PIO were in the SOPC Builder,the software programming of each module were in Nios?.Finally,on the image processing platform of FPGA,the image acquisition,image preprocessing,image edge detection and the defect classification of wood painted surface is completed.