Active Cart Detection Through Depth-Based Multi-Cart Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In retail environments, multiple shopping carts near checkout terminals complicate the identification of the active cart for checkout, leading to inefficiencies and laborious manual verification processes.
Innovation Solution
A system integrating computer vision object detection and depth estimation uses a pre-trained cart detection model and depth estimation model to analyze images from a bottom camera, generating bounding boxes and depth maps to separate active and inactive carts based on normalized depth values, reducing processor and memory usage by ignoring inactive carts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple shopping carts are monitored simultaneously at checkout terminals, then comprehensive cart monitoring coverage is improved, but cart identification accuracy deteriorates
Solution Approach 1:
The patent transitions from 2D image analysis to 3D spatial understanding by incorporating depth estimation. Depth maps provide the third dimension (z-axis distance information) that enables the system to differentiate between carts at various distances from the camera, allowing accurate identification of the active cart even when multiple carts are present in the monitoring area.
Solution Approach 2:
The patent introduces depth maps as an intermediary data layer between the raw image and the cart identification process. This intermediate depth information acts as a mediator that helps distinguish which cart is actively being used by providing distance-based differentiation, thereby improving identification accuracy without reducing monitoring coverage.
2Quantity of substance
If all carts in the field of view are processed, then complete cart analysis is improved, but processing time deteriorates
Solution Approach 1:
The patent extracts and processes only the relevant cart data by using depth information to identify and isolate the active cart from the field of view. Instead of processing all detected carts, the system extracts the specific cart that is actively being used (based on depth/distance criteria), significantly reducing processing time while maintaining complete analysis of the necessary cart.
Solution Approach 2:
The patent applies partial action by processing a subset of carts rather than all carts in the field of view. By using depth estimation to determine which cart is active, the system performs only the necessary processing on the relevant cart, avoiding the time-consuming processing of inactive carts while still achieving complete analysis of the active cart contents.
3Measurement precision
If manual cart verification is performed, then item verification accuracy is improved, but labor requirements deteriorates
Solution Approach 1:
The patent implements self-service by enabling the system to automatically identify and verify cart contents without human intervention. The combination of object detection and depth estimation allows the system to autonomously determine which cart is active and process its contents, replacing manual verification while maintaining high accuracy and eliminating labor requirements.
Solution Approach 2:
The patent substitutes the mechanical manual verification process with an automated computer vision system. Instead of human operators physically examining carts and items, the system uses algorithms to detect carts, estimate depth, identify the active cart, and verify its contents automatically, replacing the mechanical human labor with an automated computational system.
4Measurement precision
If depth estimation and object detection are integrated, then cart separation accuracy is improved, but computational resources deteriorates
Solution Approach 1:
The patent applies preliminary action by generating depth maps and detecting cart bounding boxes before performing the actual cart separation and identification. This preliminary processing organizes the data in advance, creating structured depth and location information that facilitates more efficient subsequent processing, thereby improving separation accuracy while managing computational resource usage through staged processing.
Data Source
AI summary
Examples provide for multi-cart separation using computer vision object detection and depth estimation. An image capture device associated with a checkout terminal in a retail facility generates images of shopping carts near the checkout terminal. A pre-trained cart detection model analyzes the images and detects the carts in each image. The detected carts are identified by bounding boxes in the image data of the images. A depth estimation model generates a depth map based on the image data. The depth map and bounding boxes are combined to create depth values for each cart in the images. The depth values are normalized. The normalized depth values are used to identify an active cart currently checking out at the checkout terminal in real-time. A multi-cart data set is annotated with an active cart label corresponding to the identified active cart to identify active and inactive carts in bottom camera images with greater accuracy.