Research On The Detection And Classification Method Of Daily-use Porcelain Based On Multi-agent Technology

Due to the influence of environment and material in the production process,ceramic products will appear in different degrees and different types of defects,such as cracks,spots,holes,deformation,etc.Artificial discrimination and classification can lead to low productivity.In order to improve the classification efficiency and intelligence degree of ceramic product quality grade,a kind of detection and classification method based multi-agent for daily-use porcelain is proposed in this paper,so as to realize the intelligentization and automation of porcelain detection and classification.Specific results are as follows.1.The hardware and software platform of the defect detection system of the daily porcelain ware is built,including the detection line,the illumination system,the upper computer display platform and the classification software algorithm,the scheme of the detection and classification system for daily-use porcelain based on multi agent technology is constructed,designed the function module of quality grade classification of daily porcelain.2.Digital image processing technology is used as a multi agent visual nondestructive testing technology,defect of daily-use porcelain and data analysis,detect the defects of daily-use porcelain and data analysis,the detection methods of some main defects of the domestic porcelain are obtained.Aiming at the common defects of daily bowls of several cracks,slag,a deformation,gap and baking glaze sticking test,using image processing technology to extract the boundary of the bowl,central location of bowl rim,through determine deviation from the center to determine deformation,good application effect;By increasing the contrast of linear gray scale value,using the Laplasse point detection operator to extract the internal defects to determine the slag and cracks,analyse the relationship between the various points on the boundary and the center,by means of curve radius,radius of the residual curve and radius of residual exponential curve a variety of programs given gap and patterns out of bounds criterion.3.According to the various defect detection obtained data that using the Kalman filter and C4.5 decision tree algorithm to complete porcelain quality level of classi-fication.Fir st,detected abnormal data smoothing filter process and simulation.then,the processed dat a using C4.5 decision tree algorithm to establish defectlevel tree,in accordance with the a greed rules to rank classific-ation,ignore sticky glazed porcelain such as decorated that th edefect does not affect the daily-use,subdivision quality levels.By applying the method to detect and classify ceramic products,the results show that the method in the detection and classification of porcelain obtain good application effect.This will achieve the detection and classification of ceramic products intelligent automati on,improve production efficiency of ceramic enterprises and quality of ceramic products,promote the development of China’s ceramic industry that has important theoretical signi ficance and application value.