Whole Store Scanner Using Camera Network for Automated Checkout
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional grocery shopping requires customers to manually scan items at checkout lanes, which is inefficient and can be costly to implement with smart cart and smart shelf technologies, and existing solutions often rely on mobile devices or RFID tagging.
Innovation Solution
A network of low-cost cameras with processors is used for video monitoring and tracking, analyzing events within predetermined zones to track cart and shelf contents without the need for mobile devices or special equipment, spreading processing burden across multiple camera stations and coordinating intelligence through a central system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If smart cart and smart shelf technologies are implemented, then automated item tracking is improved, but implementation cost and device complexity increase significantly
Solution Approach 1:
The patent replaces complex mechanical/electronic smart cart systems and smart shelf systems with a vision-based detection system using cameras and image processing algorithms. This substitution eliminates the need for expensive RFID tags, smart cart electronics, and complex shelf sensors, thereby reducing implementation cost while maintaining automated tracking capability
Solution Approach 2:
The system creates visual copies (images) of items and their locations using standard cameras, then processes these copies through image recognition algorithms to track items. This approach avoids the need for physical smart tags or electronic identifiers on each item, reducing overall system complexity and cost
2Measurement precision
If RFID tagging is used for item tracking, then item identification accuracy is improved, but device complexity and implementation cost increase
Solution Approach 1:
The patent substitutes RFID electromagnetic tagging systems with optical vision-based identification using standard cameras. The system captures images of items and uses computer vision algorithms to identify and track them, eliminating the need for RFID tags while achieving comparable identification accuracy through visual recognition
Solution Approach 2:
The system introduces an intermediary layer of image processing and recognition algorithms that translate visual information from standard cameras into actionable item identification data. This intermediary processing layer replaces the direct RFID reading mechanism, simplifying the overall system architecture
3Extent of automation
If centralized processing is used for video analysis, then coordination intelligence is improved, but computational burden and processing time increase
Solution Approach 1:
The patent divides the video processing task into segments, with each camera or camera group handling its own field of view independently through local processing. This segmentation allows parallel processing of multiple video streams simultaneously, reducing overall processing time while maintaining coordinated tracking through selective event communication between segments
Solution Approach 2:
The system performs partial processing at the camera level (detecting motion, identifying items of interest) and only transmits relevant events or extracted features to the central system for coordination. This partial action approach reduces the computational burden on centralized processors while maintaining effective coordination intelligence
Data Source
AI summary
Video monitoring and tracking techniques are addressed to allow consumers to purchase items in a store with no need to checkout at a traditional checkout lane. A large number of cart check and shelf check cameras monitor additions to carts and removals from shelves along with knowledge of location and what is on a particular shelf are employed to analyze which products are selected for purchase. Customer analytic data, as well as, store inventory data are preferably also developed from the camera image data.