Research On Algorithm Of Commodity Package Printing Defect Detection Based On Machine Vision

Detecting defects in commodity package printing based on machine vision has good research and application value.This topic research is aimed at inspecting whether the product package printing are qualified from the aspects of structure,color,etc.The algorithm can detect defects such as lack of printing,stains,color shift,offset,etc.,and classify genuine and defective products,aiming to serve the need of production lines and improve production efficiency.This paper focuses on the design of a defective product recognition algorithm based on template matching.Among them,the detection of the consistency of the printed structure is the core of the algorithm in this paper.This paper proposes two algorithms based on the distance between edge points and HOG features of images to solve this problem.In this paper,Hough line detection is used to estimate the initial pose of the image,and an image pyramid structure is constructed to perform layer-by-layer traversal from coarse to fine,which improves the speed of the algorithm.This traversal mechanism also ensures the scale invariance and rotation invariance of the algorithm.By selecting the appropriate image color channel,the algorithm’s lighting adaptability is enhanced.Experiments prove that the algorithm designed in this paper can effectively sort out multiple commodity packages automatically,has strong robustness and sorting efficiency,and has high application value.The algorithm system was implemented and evaluated in C language on Visual Studio platform,and the feasibility of the algorithm system was verified through experiments.The detection accuracy of the algorithm reaches more than 95%,and the detection speed reaches3 per second.