Movable Shelf Camera for High-Resolution Product Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned sales counters face challenges in accurately identifying products using weight sensors and cameras due to low resolution, increased computation, and high manufacturing costs from multiple camera installations, especially when shelves are long or products are obscured.
Innovation Solution
A system with a sensing means, camera, and driving means that moves the camera to the area of product retrieval or insertion, using weight sensors to detect changes and a moving device to capture high-resolution images, reducing the need for multiple cameras.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple cameras are installed on the unmanned sales counter to capture all shelves, then the coverage area is improved, but the manufacturing cost increases
Solution Approach 1:
The patent employs a movable camera that can dynamically change its capturing position and orientation to cover different shelf areas. The camera is mounted on a moving device that allows it to relocate along the shelf, providing comprehensive coverage over time rather than requiring multiple simultaneous cameras. This dynamic approach replaces static multiple-camera infrastructure with a single mobile imaging unit.
Solution Approach 2:
The patent introduces temporal dimension to the coverage problem by having the camera move through different positions at different times. Instead of capturing all shelves simultaneously with multiple cameras (spatial solution), the system captures different shelf portions sequentially as the camera moves, achieving complete coverage through time-based dimensionality.
2Ease of manufacture
If a single camera is used to capture all shelves, then the manufacturing cost is reduced, but the image resolution decreases
Solution Approach 1:
The movable camera system dynamically adjusts its distance and angle to each shelf section being captured, optimizing image resolution for the current field of view. By positioning the camera closer to the specific shelf area being photographed at any given moment, the system maintains high resolution images without requiring multiple cameras or wide-angle lenses that compromise detail.
3Quantity of substance
If one camera captures a wide area, then the number of cameras is reduced, but the computation required for image comparison increases
Solution Approach 1:
The patent divides the shelf monitoring task into smaller segments by having the camera focus on and capture individual shelf portions sequentially rather than capturing the entire shelf area in one wide shot. This segmentation reduces the number of products visible in each image, thereby decreasing the computational burden of comparing images before and after weight changes to identify retrieved products.
4Area of stationary object
If the camera is positioned at a certain distance to capture a wide area, then the coverage is improved, but products may be obscured by the buyer
Solution Approach 1:
The movable camera can dynamically adjust its position to maintain optimal viewing angles and distances relative to the products on the shelf. When a buyer approaches, the camera can relocate to positions that minimize obstruction, ensuring continuous reliable capture of product details regardless of buyer presence or movement.
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
Improves product identification accuracy by capturing narrow areas with high resolution, minimizing camera usage, and reducing manufacturing costs while handling various products and changing shelf configurations.
Implementation Method 1
Weight sensors are installed on the shelf where products are placed to detect the weight of the shelf. When the shelf weight changes, for instance, if the weight of the shelf decreases by 100 g, it can be determined that a product with a weight of 100 g has been retrieved by the user from the products placed on the shelf.
Implementation Method 2
The camera captures the products on the shelf, especially capturing the products on the shelf after a weight change. From the captured image, the current products on the shelf are recognized, and the retrieved and inserted products are identified by comparing with the database that records the current shelf status.
Data Source
AI summary
The unmanned sales system has a weight sensor for detecting whether a product has been retrieved from or inserted into each of plural placement areas on a shelf where products to be sold are placed, a camera for capturing the products on the shelf, a moving device for moving the camera to change its capturing area, a control unit that controls the moving device to move the camera to the placement area detected by the weight sensor, and a product identification unit that identifies the retrieved or inserted product based on the image of the placement area captured by the camera. Within the unmanned sales system, the camera moves to the area where the product retrieval or insertion is detected by the weight sensor and captures that area. Since the camera captures a narrow area, the resolution of the captured image is enhanced.