Defects Detection Of Printed Matter Based On Image Processing Research

Printed matter is one of the important carriers for people to enjoy the dissemination of spiritual and cultural information,so people have higher and higher requirements for the quality of printing.How to detect and discover printed product defects in time,reduce the influx of printing waste into the market,and reduce printing costs and human resources is a hot spot that people pay attention to.The quality defects of printed products have the characteristics of many types and complex causes.However,with the continuous development of image processing and other technologies,image processing technology has become one of the effective means of detecting defects in printed products.Aiming at the defects of existing printing quality,this paper has carried out the research of printing defects detection algorithm based on image processing.The main research contents include: firstly,in order to effectively reduce the influence of noise data on the accuracy of prediction,the paper proposes the preprocessing and edge processing methods that meet the three types of print defects of dots,lines and circles;secondly,no loss of prints Defect image features,this paper carried out a comparative analysis of five different feature extraction algorithms,Sobel operator,Roberts operator,Prewitt operator,Laplace operator and Canny operator,and proposed an improved Sobel feature extraction algorithm to obtain better edges Extraction effect;Finally,through BP network and convolutional neural network classifiers for experimental analysis and comparison,an improved BP neural network classification algorithm is proposed.Verification by actual data shows that the method proposed in this paper can not only effectively reduce the cost of printing quality inspection,but also effectively improve the effect of printed product quality inspection.The results have important promotion and application value.