Retail Display Exception Detection Using OCR and Planogram Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In retail and warehouse environments, it is challenging for humans to accurately identify misplaced or mispriced items due to the difficulty in remembering precise planogram arrangements and prices, leading to sub-optimal sales and increased costs from non-compliance, which existing methods struggle to efficiently address.
Innovation Solution
A system comprising a robot equipped with sensors to capture images and localization data, a processor to analyze images for object detection and optical character recognition, and a server to compare data against a catalog for discrepancies, generating reports on exceptions such as misplaced or mispriced items, thereby streamlining the identification process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual identification of misplaced or mispriced items is used, then human judgment can be applied, but the speed and accuracy of identification deteriorates due to difficulty in remembering planogram arrangements and prices
Solution Approach 1:
The patent replaces manual human inspection with an automated robotic system equipped with sensors, image processing algorithms, and machine learning models. The robot autonomously navigates retail environments, captures images of product displays, and uses computer vision to detect planogram compliance, thereby eliminating human memory limitations and significantly improving both accuracy and speed of identification.
Solution Approach 2:
The system enables self-service by allowing the robotic platform to independently perform data collection, image processing, exception detection, and report generation without continuous human intervention. The autonomous robot captures its own operational data, processes images locally, and generates compliance reports automatically, reducing the need for manual oversight while maintaining high identification accuracy.
2Productivity
If automated robotic systems are deployed to identify exceptions, then identification speed and accuracy improve, but device complexity increases
Solution Approach 1:
The robotic platform is designed as a multi-functional system that performs navigation, image capture, data collection, exception detection, and report generation within a single integrated device. This universal approach consolidates multiple separate systems into one platform, improving identification speed while managing complexity through functional integration rather than proliferation of separate devices.
Solution Approach 2:
The patent introduces a server as an intermediary between the robotic platform and the exception detection algorithms. The robot collects and transmits image data to the server, which then performs complex image processing and analysis. This intermediary architecture distributes computational complexity across multiple components, allowing the robot to maintain simplicity while achieving high identification speed through centralized processing.
3Measurement precision
If comprehensive image analysis is performed on all product displays, then exception detection accuracy improves, but the time required for processing increases
Solution Approach 1:
The system employs selective image processing by focusing analysis only on regions containing potential exceptions rather than processing entire product displays uniformly. The robotic platform identifies areas of interest through preliminary scanning and applies detailed analysis only to those specific regions, thereby maintaining high exception detection accuracy while significantly reducing overall processing time through targeted rather than comprehensive analysis.
Data Source
AI summary
Systems and methods for identifying exceptions in feature detection analytics include systems and methods configured to identify exceptions in product displays. In one exemplary embodiment, price tag mismatches are reported to a customer if certain criteria are met based on analytics from optical character recognition, image object detection, and reference catalogs/planograms. These price tag mismatches, for example, provide actionable insights for humans working alongside robots which can be quickly identified and resolved.