Supermarket intelligent vending system

A supermarket, intelligent technology, applied in the direction of image data processing, equipment, sale/lease transactions, etc., can solve problems such as congestion, shoppers queuing, etc.

Inactive Publication Date: 2019-07-05
SHANGHAI MARITIME UNIVERSITY
View PDF0 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In the traditional supermarket mode, the checkout is carried out by manual scanning of goods. This mode is very easy to cause congestion, resulting in a large number of shoppers queuing at the checkout. The entire checkout process is limited by the space at the cashier and the number of cashiers. Due

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Supermarket intelligent vending system
  • Supermarket intelligent vending system
  • Supermarket intelligent vending system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0123] This embodiment realizes a process of parameter initialization of a supermarket intelligent vending system.

[0124] 1. The image preprocessing module does not work during the initialization phase;

[0125] 2. In the initialization process of the human target detection module, the parameters of the target detection algorithm are initialized using the images with the calibrated human body image area, face area, hand area and product area.

[0126] The use of the image with the calibrated human body image area, human face area, hand area and product area to initialize the parameters of the target detection algorithm is as follows: the first step is to construct a deep network for feature extraction; the second step , Construct the area selection network, the third step, extract each image X in the database used in the deep network and the corresponding human body area marked manually according to the constructed feature extraction Then through the ROI layer whose input ...

Embodiment 2

[0178] This embodiment implements a detection process of a supermarket intelligent vending system.

[0179]1. The image preprocessing module, in the detection process: the first step is to denoise the mean value of the surveillance image taken by the surveillance camera to obtain the denoised surveillance image; the second step is to denoise the surveillance image after denoising Light compensation, so as to obtain the light-compensated image; the third step is to perform image enhancement on the light-compensated image, and transfer the image-enhanced data to the target detection module.

[0180] The monitoring image taken by the monitoring camera carries out mean value denoising, and its method is: the monitoring image taken by the monitoring camera is set as X src , because X src is a color RGB image, so there is X src-R , X src-G , X src-B Three components, for each component X src ’, perform the following operations respectively: first set a 3×3-dimensional window, c...

Embodiment 3

[0243] This embodiment implements a process of updating a product list in a supermarket smart vending system.

[0244] 1. This process only uses the product identification module. When changing the product list: if a product is deleted, the image of the product will be deleted from the product image collection of each angle, and the corresponding position in the product list will be deleted; if a product is added, the product image of each angle of the current product will be deleted Put in the collection of product images from various angles, add the name of the currently added product after the last digit of the product list, and then update the product recognition classifier with the new collection of product images from various angles and the new product list.

[0245] The method for updating the product recognition classifier with new product image collections from various angles and a new product list is as follows: the first step, modifying the network structure: for th...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a supermarket intelligent vending system which aims to solve the problem that the manual commodity scanning is too time-consuming in the traditional checkout process, and the commodity input process is advanced to a shopper to take commodities, so that the time consumption for scanning the commodities in the checkout process is reduced, the checkout speed is greatly increased, and the shopping experience of customers is improved. According to the invention, a mode recognition algorithm is utilized to recognize and count actions in a shopper goods selection process, by identifying a picture of a commodity when a customer takes and places the commodity, a commodity type is obtained, the face identification is carried out on the customer, and the human body image identification is used to obtain the identity of the customer when the face identification is not ideal, and by identifying an abnormal behavior of the customer, whether a theft behavior exists or not is judged. The system can realize an automatic statistics function on the premise that the shopping experience of customers is not reduced. The invention relates to a shopping accounting process of customers without changing an original organization structure of a supermarket, thereby facilitating the seamless joint with an existing supermarket organization structure.

Description

technical field [0001] The invention relates to the technical field of computer vision monitoring, the field of target detection, target tracking and pattern recognition, and in particular to the field of detecting, tracking and action recognition of individuals in front of shelves based on monitoring cameras. Background technique [0002] In the traditional supermarket mode, the checkout is carried out by manual scanning of goods. This mode is very easy to cause congestion, resulting in a large number of shoppers queuing at the checkout. The entire checkout process is limited by the space at the cashier and the number of cashiers. Due to the limitations of the traditional cashier model, checkout congestion is unavoidable; although existing customers can reduce the time spent on scanning products by scanning their own products for checkout, they still need to manually check the products when they go out, which will still cause congestion. Analyzing the cause of the congestio...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00G06K9/62G06Q30/06G06T5/00G06T7/90
CPCG06T5/002G06T7/90G06Q30/06G06V40/16G06V40/107G06F18/214
Inventor 刘昱昊
Owner SHANGHAI MARITIME UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap