Application Of Image Processing In Color Printing Quality Inspection

With the progress of printing technology,printing from black and white to color,and how to carry on the on-line detection of color printing has become a problem for the printing enterprise.In the actual printing production,due to the influence of printing process and printing machinery accuracy,printing always has all kinds of defects,such as omit,stain,blur,color distortion and so on.Therefore,search of printing quality automatic inspection based on machine vision and image processing can effectively solve the problem of color printed online detection.This paper mainly research on the following three aspects:1.Printing will offset and scale due to the influence of mechanical shock and paper stretch in the process of transmission and image acquisition.The accuracy of the detection work will be affected because the collected printing image and the standard image do not match in the spatial position.Therefore,aiming at the problem of color printing registration,the registration algorithm based on saliency region is put forward here.Traditional SURF algorithm is changed to a new SURF algorithm by adding color features in SURF feature descriptor;then the saliency region of image is extracted by the saliency algorithm based on super-pixel;the new SURF algorithm is used to detect and match the feature points in the image saliency region to realise image registration.2.The purpose of detecting printing product defects is to be able to locate and mark of printing defects quickly and accurately,enables the staff determined the types of defects quickly,and take measures to solve the problems which caused printing defects.The two images are divided after the image registration,and then the automatic threshold processing is performed to get the binary image.Marking the location of the defect by tracking the edge of the binary image.3.For the chromatic aberration of color printing,in order to make an accurate and efficient judgment,it is necessary to preprocess the collected image and standard image before calculate the chromatic aberration.This topic introduces the super pixel in image preprocessing stage,the image is segmented into many super-pixels which are some compact structure and approximate uniform pixel block by gathering adjacent pixels with similar characteristics together.Each super-pixel replace the multiple similar pixels in pixel block to calculate.In order to prove the effectiveness of the above three methods,experiments are carried out respectively.The experimental results show that the proposed method is faster than extracted the point of whole image,is more precise than randomly selected sub region;defect detection method for the automatic threshold processing can get the best two binary images according to the different defects in the color printing;the chromatic aberration method for color printing can greatly improve the efficiency of online detection on the basis of ensuring the detection accuracy.