Unknown Device Registration Using AI Image Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fleet management systems struggle with registering devices from unknown brands or those lacking standard identification, leading to incomplete registration and manual processes that are error-prone and time-consuming, resulting in an incomplete view of the fleet.
Innovation Solution
A system and method utilizing Artificial Intelligence (AI) based image recognition to extract object identifiers from unknown objects, transmitting them to external sources for refined information using a Large Language Model (LLM) to automate the registration process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual registration process is used for unknown devices, then registration can be performed, but it is error prone and time consuming
Solution Approach 1:
The patent replaces the manual mechanical registration process with an automated optical system using image capture devices and AI-based image recognition algorithms. The system automatically extracts object identifiers from images of unknown devices, eliminating the need for manual data entry and significantly improving registration efficiency while maintaining accuracy.
2Loss of information
If barcode decoding is used for device registration, then identification can be obtained, but decoding and interpreting barcodes may lead to registration failures
Solution Approach 1:
The patent introduces an intermediary AI-based image recognition system that acts as a mediator between the physical device and the registration system. Instead of directly relying on barcode decoding which may fail, the system captures images of devices, uses AI to extract various types of identifiers (text, serial numbers, manufacturer information), and processes this information through multiple steps to achieve reliable registration.
3Device complexity
If only branded devices are registered in the FLM system, then registration process is simplified, but the view of the entire fleet becomes incomplete
Solution Approach 1:
The patent changes the key parameter for device identification from requiring standardized brand information to accepting any extractable identifier from device images. The AI system is designed to recognize and extract various types of information including text elements, serial numbers, manufacturer details, and device type labels from images of both branded and unbranded devices, thereby expanding fleet registration completeness without significantly increasing process complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of the present disclosure provide a system (200) and method (300) for registering unknown objects into a device management system. The method comprises receiving (302), through a processor (204), one or more object data related to an object. The object is unknown to the device management system. The method further comprises extracting (304) one or more object identifiers from the one or more object data to identify the one or more objects, transmitting (306) the one or more object identifiers to an external source to identify a refined information about the object. The refined information is used to classify the object. The method further comprises registering (308) the object with the device management system according to the classification of the object.