Design And Implementation Of Pharmacy Inventory Management System Based On Image Recognition

With the development of automated pharmacies,more and more automated equipments are applied to the pharmacy system,but current equipments are only used in the pharmacy's dispensing system.The work of tonics still needs to be done manually,which greatly affects the work efficiency of pharmacy.Through the discussion of pharmacy automation equipment and automatic tonic process,an inventory management system based on image recognition is designed,which mainly completed the automatic warehousing of drugs and the inventory management of pharmacy drugs,solved the problem of pharmacy tonic automation,and improved the automation level of pharmacies.This paper analyzed the requirements of the system and designed the overall structure of the system.In this paper,the hardware structure of the system is designed and the selection of the PLC and motors are completed.The communication protocol between the upper and lower machines is established,and the PLC program is written.The system database is designed to complete the recording of drug information and staff workload information.According to the characteristics of the kit,the identification procedure of the kit is designed.The Canny operator is used to detect the edge of the kit,and the Hough line detection algorithm is combined to propose the segmentation method of the kit.The corrosion expansion operation and the MSER algorithm are used to detect the text block on the surface of the cartridge,and we use OCR to identify processed blocks of text to determine the name of the drug.In order to prevent the drug name identification from being wrong,the method for calculating similarity of drug name string is proposed based on the editing distance of the Chinese character string and the similarity of the single Chinese character.The method output the name of the drug with the highest similarity in the database as the recognition result.An auxiliary recognition algorithm based on the size of the pill box and the surface color of the pill box is designed to verify the accuracy of the main algorithm identification and further improve the accuracy of the overall recognition algorithm.The system operation interface is designed,and the overall debugging work of the system is completed,which verified the feasibility of the system.