Research And Design Of Unstaffed Convenience Store Commodity Purchase System Based On Machine Vision

The unstaffed convenience store is one of the classic applications in the unmanned scene,and it is a new retail market with great potential.The convenience store is operated well based on the commodity purchase system,which can meet the whole consumption process without staff participation.However,there is no effective balance between the consumption experience and the cost in the current market for the purchase system of convenience store goods.In view of the above problems,this paper studies and proposes a machine vision based commodity purchase system,which can help the convenience store to make statistics of customers' consumption and generate corresponding bills.The main contents of this paper are as follows:(1)First of all,the system level design part is based on the actual needs of the convenience store,combined with the actual operation needs of the convenience store under the offline scene.Using the idea of function refinement and decomposition,the commodity purchase system is divided and designed,which is named commodity purchase subsystem,commodity purchase secondary verification system and inventory update system respectively.Among them,the main function of the commodity purchase subsystem is to count the consumption of customers entering the store and generate corresponding bills.The main function of the secondary verification system of commodity purchase is to verify the statistical results of customer consumption generated by the subsystem of commodity purchase and explore its rationality.The inventory update system counts consumption results and updates shelf goods in real time,automatically prompts when inventory is insufficient.(2)Secondly,in the part of hardware and algorithm design according to the functional requirements of different subsystems,the design concepts and ideas of the commodity purchase subsystem,the commodity purchase secondary verification system,and the inventory update system are introduced respectively.Among them,the functional requirements and non functional requirements are analyzed,the corresponding hardware part and the core algorithm part are designed.The effectiveness and stability of the algorithm are proved by offline test analysis,comparison and verification.(3)Finally,in the part of the offline entity system test and later work development of the convenience store,the offline scene data is used to test the system and find the later optimization direction.Among them,the practicability and robustness of each subsystem and the whole system are verified by massive data and entity operation.It has been proved that the system can run normally and stably effectively,meet the needs of the actual scene,and has good engineering application value.After passing the test,analyze the results and put forward the later optimization direction of the algorithm.