Neural Network Device Identification for Inspection Automation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Inspectors and maintenance personnel face inefficiencies in locating and retrieving information for various devices due to lengthy lists, especially on mobile devices, making it time-consuming and difficult to identify and access relevant checklists and procedures.

Innovation Solution

A method involving capturing images of devices using a mobile device's camera, transmitting these images to a server with a neural network for identification, and retrieving and displaying associated checklists, which includes tasks or questions for inspection processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If workers browse long lists of devices to locate inspected devices, then they can access device information, but the process becomes time-consuming and inconvenient

Engineering Contradiction:

Improvetime to locate device informationVSAvoidease of browsing device list

Core Design Contradiction:

Loss of timeVSEase of operation

Solution Approach 1:

The patent replaces the mechanical manual browsing system with an automated neural network image recognition system. Instead of workers manually scrolling through long device lists on mobile devices, the system captures images of devices and uses neural networks to automatically identify and retrieve associated information, eliminating the need for manual list browsing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by allowing workers to simply capture an image of the device they need information about, and the system automatically handles the entire information retrieval process without requiring workers to manually search through databases or lists

Inventive Principle:

Principle #25Self-service

2Loss of information

If workers manually search through lengthy device lists, then they can find inspected devices, but the process becomes difficult on mobile devices

Engineering Contradiction:

Improveaccess to device informationVSAvoidease of locating device in list

Core Design Contradiction:

Loss of informationVSEase of operation

Solution Approach 1:

The patent replaces the mechanical manual search process with automated image-based identification. Workers capture images of devices using mobile devices, and the neural network automatically processes these images to retrieve device information, eliminating the difficulty of manually searching through lengthy lists on mobile devices

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces an intermediary - the neural network image recognition system - that mediates between the worker and the device information database. Instead of workers directly searching through lists, they simply capture images, and the neural network intermediary automatically translates these images into device identification and information retrieval

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11688051B2Automating inspection using neural network

Publication Date: 2023.06.27 TYCO FIRE & SECURITY GMBH

AI summary

Aspects of the present disclosure include methods, systems, and non-transitory computer readable media that perform the steps of capturing capture one or more images comprising a plurality of visual features of an inspected device, transmitting the one or more images to a server comprising a neural network, wherein the neural network analyzes the plurality of visual features to identify the inspected device and the server identifies the checklist comprising tasks or questions associated with an inspection process of the inspected device, receiving the checklist associated from the server, and displaying the checklist.