Optical Flow Object Recognition for Shopping Cart Merchandise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for detecting and recognizing objects on the lower shelf of shopping carts in retail environments are prone to false positives, lack integration with existing checkout systems, and struggle with efficient image selection for object recognition, leading to productivity issues and unaddressed collusion.
Innovation Solution
A system that uses optical flow to select and process images by capturing and partitioning images into blocks, comparing them for displacement, and identifying common displacement measurements to isolate cart movement, allowing for efficient feature extraction and recognition of items on the cart.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional infrared sensors are used to detect merchandise presence, then the system can detect object presence, but it cannot identify object identity and produces false positives
Solution Approach 1:
The patent replaces conventional infrared sensors with a video camera system that captures visual images. This substitution enables the system to not only detect object presence but also capture visual information for identification, thereby eliminating false positives while providing both presence detection and identity recognition capabilities.
Solution Approach 2:
The patent introduces an image selection algorithm as an intermediary between the video camera and object recognition system. This intermediary selectively processes only relevant frames containing merchandise, reducing computational load while preserving object identity information for accurate recognition and integration with the POS system.
2Reliability
If video surveillance is used to monitor the bottom of the basket, then object presence can be visualized, but the system is not integrated with POS and relies on cashier intervention
Solution Approach 1:
The patent implements an automated system where the video camera captures images, the image selection algorithm processes frames, and the object recognition system automatically identifies merchandise. This self-service automation eliminates the need for cashier intervention in detecting and identifying bottom-of-basket items, thereby maintaining high reliability while significantly improving checkout productivity.
Solution Approach 2:
The patent creates a multi-functional system that combines video surveillance capabilities with POS system integration. The same image processing pipeline serves both visual monitoring and automated merchandise identification functions, enabling the system to simultaneously provide reliable detection and automated checkout processing without requiring separate systems.
3Measurement precision
If all captured images are processed for object recognition, then complete object identification is achieved, but processing capacity is exceeded and timeliness is lost
Solution Approach 1:
The patent applies partial action by processing only a subset of captured images rather than all frames. The image selection algorithm identifies and processes only those frames that contain merchandise of interest, discarding redundant frames. This selective processing maintains complete object recognition accuracy for relevant items while dramatically reducing the total processing time and computational load.
Data Source
AI summary
Disclosed is a system and method for using optical flow to detect objects moving past a camera and to select images of the moving objects. A shopping cart, for example, may be detected by subdividing an image into a plurality of image blocks; comparing the blocks to a preceding image to determine the motion of the portion of the object pictured; associating the most common motion with the shopping cart. The motion of the cart may also be integrated over time for purposes of tracking cart motion and selecting a subset of the captured images for object recognition processing. Detection of the cart and image selection improves computational efficiency an increase merchandise throughput.