Store Shelf Imaging with Overlap Correction for Product Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Retail chains face challenges in automatically and accurately documenting product locations on shelves across stores due to variations in product packaging, orientation, and lighting conditions, leading to inefficient manual sorting and signage placement.
Innovation Solution
A robotic system with a mobile base, image capture devices, and a master control unit that captures sequential overlapping images to perform image-based accuracy correction, determining the optimal shift for aligning images and correcting the mobile base's location, enabling automated product location mapping and signage packaging in the correct order.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual sorting of signage is performed to match store product locations, then signage can be correctly placed, but time consumption increases significantly
Solution Approach 1:
The patent replaces manual mechanical sorting operations with an automated image processing system. The system captures images of product locations, processes them through algorithms to identify and map products automatically, and generates signage packages in the correct order without manual intervention. This substitution of mechanical/manual processes with automated computational processes directly resolves the contradiction by maintaining accuracy while eliminating time-consuming manual sorting.
Solution Approach 2:
The system enables self-service automation where the image processing system independently performs the complete workflow of capturing store images, identifying products, mapping locations, and generating signage packages in the correct sequence. The system serves itself by automatically correcting location data and ordering signage without requiring manual sorting operations, thereby resolving the time-accuracy contradiction.
2Loss of time
If automated image processing is used to document product locations, then time is reduced, but accuracy decreases due to variations in packaging, orientation, and lighting
Solution Approach 1:
The patent applies preliminary action by implementing a multi-stage image processing pipeline that performs preliminary corrections for lighting variations, packaging differences, and orientation issues before final product identification. The system pre-processes images to normalize these variations, then proceeds with accurate product recognition. This preliminary correction of potential errors before the main identification task resolves the contradiction by maintaining high accuracy while keeping the process automated and time-efficient.
Solution Approach 2:
The system incorporates feedback mechanisms where the image processing algorithm continuously refines its product identification based on detected variations in packaging, orientation, and lighting conditions. The system uses feedback from image analysis to adjust its recognition parameters and improve accuracy dynamically, thereby maintaining precise product identification throughout the automated process without sacrificing speed.
3Area of stationary object
If multiple image capture devices are used to improve coverage, then more products are captured, but system complexity increases
Solution Approach 1:
The patent applies merging by combining multiple image capture devices into a unified system where images from multiple sources are integrated and processed together. The system merges the captured images, aligns them based on their spatial relationships, and processes them as a cohesive dataset. This merging approach allows the system to maintain wide coverage from multiple devices while managing complexity through unified processing algorithms rather than treating each device independently.
Solution Approach 2:
The image processing system is designed with universality to handle inputs from multiple different image capture devices with varying specifications. The system can process images from different devices, orientations, and resolutions using a single unified algorithmic framework. This multi-functional capability allows the system to leverage multiple devices for increased coverage without proportionally increasing system complexity, as the same processing pipeline handles all inputs.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A store profile generation system includes a mobile base and an image capture assembly mounted on the base. The assembly includes at least one image capture device for acquiring images of product display units in a retail environment. A control unit acquires the images captured by the at least one image capture device at a sequence of locations of the mobile base in the retail environment. The control unit extracts product-related data from the acquired images and generates a store profile indicating locations of products and their associated tags throughout the retail environment, based on the extracted product-related data. The store profile can be used for generating new product labels for a sale in an appropriate order for a person to match to the appropriate locations in a single pass through the store.