Research And Implementation Of Label Detection System Based On Improved Structural Similarity

With the improvement of people's living standard,people pay more and more attention to the quality of food packaging.As an important form of packaging,labels are widely used in food production because they have the function of marking the key information of food.However,in the production process,because of the production environment,human error and other reasons,there may be some problems during the sticking processof labels,such as label damage,missing sticking and so on.This will bring losses to the enterprises and consumers.Therefore,it is necessary to detect the label.In this paper,we have studied and implemented a set of label defect detection system,which can detect the defect of the surface label attached to the container bottle.The system can also be applied to the detection of the flat label.In the process of system implementation,the algorithms of edge detection,surface correction,image feature point matching and image quality evaluation are applied.The experiment proves that the common label defect type can be detected successfully.The main work of this paper is:·Label image preprocessing design,including label area segmentation and image surface correction.After analyzing and studying the feature of the image,this paper realizes the extraction of the label area through the image difference and edge detection,and stretches the surface image through the cylindrical back projection algorithm.·In order to align the reference image and the image to be measured in space,this paper study image registration theory.And the image registration algorithm based on SIFT feature point matching is designed to minimize the spatial error of the reference label image and the pending image.·Label image defect detection algorithm design.Aiming at the characteristics and problems of the label image after preprocessing and registration,the defect detection ideas based on the image quality assessment algorithm are determined.After analyzing the feasibility and deficiencies of the common image quality assessment algorithm in the requirements of this paper,based on the structural similarity.Degree of improvement,and based on the improvement of the algorithm design flaw detection program,can be well adapted to the characteristics of the label to meet the detection needs.·On the basis of the previous work,the technical implementation of the label defect detection system is proposed,including the hardware and software implementation,and the accuracy and real-time performance of the system are tested.The test requirement shows that the defect detection system designed in this paper can meet the actual requirements.