Research On The Proofreading Technology And Application Of Irregular Cigarette Packets Based On Machine Vision

The researches and applications of industrial robots for specific industries are important contents for robots to promote industrial transformation and upgrading in manufacturing industry.In this paper,aiming at the problems of error-prone and inefficiency in manual stacking of irregular packages in traditional tobacco logistics industry,the research and application,which combined stacking of irregular packages and visual proofreading technology have important theoretical significance and practical application value.Firstly,the characteristic and process of irregular stacking of cigarette bags in tobacco logistics are analyzed.Combined with machine vision technology,the overall design of the system is completed,and the control scheme and hardware equipment of visual proofreading are analyzed and selected.The research focuses on the control system which combined stacking and the visual proofreading algorithm.In the research of the combined palletizing control system,the design of the combined palletizing control system is completed.The control function module is modularized and the corresponding hardware equipment is designed.Then the bottom PLC program and the upper computer control program are compiled,and the research platform of the control system is constructed.To solve the problem of cigarette pack combination matching,an improved time-optimal pre-scheduling matching algorithm is proposed.By increasing multiple matching methods,the number of packs in buffer area and the number of stacking paths of stacking robot are reduced,and the optimal matching method is obtained.Experiments show that this method effectively reduces the stacking time and improves the stacking efficiency.In the research of visual proofreading algorithm,aiming at the consistency proofreading problem between stacked cigarette packs and cigarette packs in the PC database,an improved ORB-SURF and RANSAC algorithm is proposed to complete the identification and proofreading of cigarette packs.This algorithm uses SURF to detect feature points quickly,and ORB descriptors to describe them,which makes up for the lack of rotation invariance of ORB algorithm.Aiming at the problem that the quality of feature points is not considered in the process of eliminating mismatched point pairs in traditional RANSAC algorithm,which results in too much computation,the sample set of feature points sampled by RANSAC is screened first,and then the neighbor matching and bidirectional intersection are adopted.In order to reduce the number of iterations and improve the matching accuracy,the cross-matching method eliminates some feature points with obvious errors.Experiments show that the improved algorithm improves the matching accuracy of irregular cigarette packs in visual alignment and has good robustness.Finally,the whole system is debugged and analyzed on site.The experimental data show that the average recognition time of the system is 0.9s,the correct recognition rate is more than 95%,and the stacking error rate is less than 0.06%.The system can efficiently complete the function of checking and stacking cigarette packs.The whole system has high stability and real-time performance for handling the stacking problem of irregular cigarette packs.