Retail Inventory Tracking via Robotic Shelf Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current inventory management systems in retail stores lack efficiency in detecting understock conditions on customer-facing shelves and fail to automatically prompt restocking from top-shelf inventory or back-of-store inventory, leading to inefficient restocking processes.
Innovation Solution
A method and system that uses a robotic system to capture images of inventory structures, detect top shelves and product units, and generate prompts for restocking understocked slots by transferring product units from top shelves to customer-facing slots or retrieving from back-of-store inventory based on image analysis and inventory data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual inventory checking and restocking is used, then operational simplicity is maintained, but productivity and detection accuracy deteriorate
Solution Approach 1:
The system enables self-service inventory management by automatically detecting understock conditions through image analysis and generating restocking prompts without human intervention. The robotic system autonomously navigates shelves, captures images, identifies product units, detects understock conditions, and creates restocking tasks, allowing the inventory system to serve itself.
Solution Approach 2:
The patent replaces manual mechanical inventory checking with an automated robotic system that uses computer vision and image processing. The robotic system substitutes human operators by using cameras to capture shelf images, algorithms to analyze product presence, and automated prompt generation to trigger restocking actions.
2Measurement precision
If automated robotic systems are deployed, then productivity and detection accuracy improve, but device complexity increases
Solution Approach 1:
The system creates a digital copy of the physical inventory state by capturing images of shelves and using image processing to generate virtual representations of product locations and quantities. This digital twin allows for accurate detection of understock conditions without physically handling products.
Solution Approach 2:
The patent introduces an intermediary image processing layer between the physical inventory and the detection system. Instead of directly sensing product quantities, the system uses captured images as an intermediary medium, which are then analyzed by algorithms to detect understock conditions accurately.
3Speed
If real-time image analysis is performed, then detection speed and productivity improve, but use of energy increases
Solution Approach 1:
The system performs partial image analysis by focusing only on relevant regions of shelves where understock conditions are likely to occur. Instead of analyzing every pixel in entire shelf images, the system identifies and processes only the critical areas containing product units, reducing computational energy while maintaining detection speed.
Data Source
AI summary
One variation of a method for tracking and maintaining inventory in a store includes: accessing a image of an inventory structure in the store; identifying a top shelf, in the inventory structure, depicted in the image; identifying a set of product units occupying the top shelf based on features detected in the image; identifying a second shelf, in the set of shelves in the inventory structure, depicted in the image, the second shelf arranged below the top shelf in the inventory structure; based on features detected in the image, detecting an understock condition at a slot-assigned to a product type-on the second shelf; and, in response to the set of product units comprising a product unit of the product type, generating a prompt to transfer the product unit of the product type from the top shelf into the slot on the second shelf at the inventory structure.