Checkout Item Counting with 3D Depth Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Checkouts are time-intensive due to the need to determine the quantity of items in bags, which can lead to errors and shrinkage due to customer inaccuracies, and conventional imaging methods fail to provide clear visibility or accurate counting, especially with non-transparent bags.
Innovation Solution
Utilizing a depth sensing or infrared camera to create a 3D map of items on a scale, projecting a plane onto the map, and counting intersections with blobs representing items to determine quantity, with verification against customer input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional imaging methods are used to count items in bags, then the system is simple and easy to operate, but the measurement precision and reliability of item counting are insufficient
Solution Approach 1:
The patent transitions from conventional 2D imaging to 3D depth mapping to improve item counting accuracy. The depth sensor captures three-dimensional information about items in bags, allowing the system to distinguish individual items based on their spatial positions and depths, thereby resolving overlapping items that conventional 2D cameras cannot separate.
Solution Approach 2:
The patent replaces manual item counting (mechanical/human operation) with automated depth sensing and computational algorithms. The system uses depth sensors and processing algorithms to automatically identify, separate, and count items, eliminating the need for customers to manually count items in their bags.
2Productivity
If customers manually count items in bags during checkout, then the system requires minimal equipment, but the transaction time increases and errors occur
Solution Approach 1:
The system performs automated item counting without requiring customer intervention. The depth sensor and processing algorithms independently identify and count items in customers' bags, eliminating the time customers would spend manually counting while improving accuracy and preventing errors.
Solution Approach 2:
The system performs item counting automatically as items are placed in bags during shopping, before the checkout process begins. This preliminary automated counting eliminates the need for time-consuming manual verification at the register, thereby increasing transaction throughput.
3Reliability
If customers enter item quantities manually at self-checkout, then the system is simple to operate, but shrinkage occurs due to lying or mistakes
Solution Approach 1:
The system provides automated feedback by comparing the depth-sensor-determined item count with the quantity entered by the customer or cashier. When discrepancies are detected, the system alerts the user and requires verification, thereby preventing shrinkage due to intentional lying or unintentional mistakes while maintaining ease of operation through automated verification.
4Object-affected harmful factors
If non-transparent bags are used to protect produce, then product protection is improved, but item visibility and counting accuracy deteriorate
Solution Approach 1:
The patent uses depth sensing to overcome the opacity of non-transparent bags. By capturing three-dimensional depth information, the system can identify the shape, size, and spatial position of individual items within the bag, allowing accurate counting even when items are not visible to the human eye through the bag material.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately counts items in bags, reducing transaction time and errors, and preventing fraud by ensuring item quantity verification matches customer input.
Implementation Method 1
obtaining depth information for a three-dimensional (3D) image depicting items on a scale at a terminal during a transaction... identifying the depth information as infrared (IR) data captured in the 3D image by an IR sensor of the terminal
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
A depth image of items on a scale are captured. Depth information for the depth image is processed to produce a three-dimensional (3D) map of the depth information. The items are identified as blobs within the 3D map. A plane is generated and projected into the 3D map and each separate blob that intersects the plane is counted to determine an item quantity for the items depicted in the original depth image. The item quantity is provided back to a terminal for item quantity verification of the items during a checkout transaction at the terminal. In an embodiment, the plane is projected onto the 3D map and animated to move from a top of the 3D map to a bottom of the 3D map while the separate blobs are counted for the item quantity determination as the plane moves.