Research And Implementation Of An Online Intelligent Printing System

This paper mainly introduces the requirements and design,overall architecture design and test process of an online intelligent printing system.The system uses the O2O e-commerce mode,connects the offline printing shop to the online through internet technology,and realizes the online printing process through WeChat payment,WeChat login,front-end and back-end Web technology,and express delivery.The front-end and back-end are divided into user’s mobile WeChat end,user’s computer end,business’s computer end,HTTP server,WebSocket server and image server,they use the technology of Tornado,Vue.js,MySQL and Redis.Through the interaction between the front-end and the back-end,the physical printing shop and the self-service printing terminal are connected to realize the whole process of online printing on the mobile WeChat end and the computer end,and the intelligent printing system is realized through the recommendation system,the modified Laplace algorithm and the artificial intelligence technology.This paper analyzes the low definition of the existing Laplacian algorithm in the calculation of the virtual or blank image,and proposes an improved Laplacian algorithm,which can improve the definition of the virtual or blank image by 2 to 4 times.The more the blank,the larger the correction range,and use deep learning technology to enhance the unclear pictures,so as to better meet the requirements of photo printing.This paper introduces User-based and Item-based recommendation algorithms and cosine similarity,modified cosine similarity and Pearson similarity.The recommender system uses a hybrid recommender system based on time and distance to recommend the print shop,constructs a minus exponential function through time and distance,adds the influence factor of time and distance into Pearson similarity,and considers the influence of the size of the intersection between users or items on the similarity,uses the size of the intersection of users or items to weight the similarity.K-NN algorithm is used in recommender system,and User-based and Item-based recommender systems are mixed by weighting time and distance.Compared with the common User-based algorithm,the mean absolute error MAE is reduced by 0.0196 when the number of neighbors is the best.This paper also proposes a WeChat login method based on WebSocket Protocol.The temporary QR code of WeChat is used to obtain the OpenID of the code scanning user and the corresponding WebSocket link,and then the Cookie string made by OpenID is issued through the WebSocket link,so as to bypass the WeChat login interface and make the login process more controllable and easier to develop.Through the existing WebSocket module,the WeChat payment process using WebSocket module to notify the payment result is designed.WebSocket active notification replaces the browser passive polling,which avoids the waste of resources and ensures the real-time notification.