Goods-selling method and device based on image comparison and self-service vending machine

Abstract

The invention discloses a goods-selling method and device based on image comparison and a self-service vending machine. The method includes the steps that in image obtaining, when a door of the self-service vending machine is opened or will be opened, images of goods on goods shelves in the self-service vending machine are obtained as original images, and after the door of the self-service vendingmachine is opened, the images of the goods on the goods shelves are obtained at preset time intervals; in goods detection, each obtained image and the previous image of the image are subjected to feature comparison to determine goods which are taken from the goods shelves by users or goods which are returned to the goods shelves; in purchase account settling, when the door of the self-service vending machine is closed or after the door of the self-service vending machine is closed, the goods which are finally taken are determined, and account settling is performed. Through the method, the images of the goods are obtained at the set time intervals; by analyzing and judging the adjacent images, the goods which are taken from the goods shelves or returned to the goods shelves by the users can be judged correctly.

Description

technical field [0001] The present application relates to the field of image processing, in particular to a vending method, device and unmanned vending machine based on image comparison. Background technique [0002] In the existing intelligent unmanned vending machines, the buyer needs the system to automatically identify the goods that the user takes away when shopping. Setting up such machines in public places not only facilitates users, but also saves a lot of labor costs. However, usually such smart vending machines use high-speed cameras to capture the process of users taking goods in real time to identify purchased goods. However, sometimes the user picks up the item too fast, which will cause the camera to be too late to track, resulting in recognition difficulties, or no recognition at all; sometimes the user holds the item completely or only reveals a part of the item, or the user intentionally or unintentionally covers the part of the item Key parts can also cau...

AI Technical Summary

Problems solved by technology

However, sometimes the user picks up the item too fast, which will cause the camera to be too late to track, resulting in recognition difficulties, or no recognition at all; sometimes the user holds the item completely or only reveals a part of the item, or the user intentionally or unintentionally covers the part of the item Key parts can also cause difficulty in identification, or can not be identified at all

All of the above situations will lead to settlement errors, making it impossible for users to shop smoothly and deteriorating user experience.

Images