Textile Machinery Inspection With Abnormality-Based Image Storage

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing inspection systems for textile machinery face challenges in managing large volumes of captured images, leading to storage capacity issues while failing to maintain an overview of the distribution of abnormality across regions, as seen in Japanese Patent Application Publications 2020-196972 and 2021-144000.

Innovation Solution

An inspection system for textile machinery that includes a classification unit to detect and group images based on abnormality scores, a counting unit to track image counts per group, and a determination unit to manage storage by replacing images when limits are reached, ensuring efficient data storage and distribution analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If all captured images are stored in the storage unit, then the complete distribution of abnormality can be obtained, but the storage capacity is strained and data management becomes difficult

Engineering Contradiction:
Improvedistribution of abnormalityVSAvoidamount of data
Core Design Contradiction:
Loss of informationVSQuantity of substance

Solution Approach 1:

The patent segments captured images into multiple groups based on abnormality degrees (e.g., group A for high abnormality, group B for medium abnormality, group C for low abnormality). This segmentation allows selective storage of only necessary images while maintaining distribution information, resolving the contradiction between complete data retention and storage capacity constraints.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and stores only the distribution information (counts of images per group) rather than all captured images. By taking out only the essential statistical data about abnormality distribution, the system reduces storage requirements while preserving the ability to analyze abnormality patterns across different groups.

Inventive Principle:
Principle #2Taking out (Extraction)

2Quantity of substance

If an upper limit on the number of captured images to be saved is set, then storage capacity is reduced, but the overall distribution of captured images across all regions cannot be obtained

Engineering Contradiction:
Improveamount of dataVSAvoiddistribution of abnormality
Core Design Contradiction:
Quantity of substanceVSLoss of information

Solution Approach 1:

By dividing images into abnormality-based groups, the system can apply different storage strategies to each group. The classification unit segments images before storage, enabling selective retention that maintains distribution information while reducing overall data volume.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The counting unit continuously monitors and stores the number of captured images in each group, providing feedback on distribution. This feedback mechanism allows the system to understand abnormality distribution patterns without storing all individual images, resolving the contradiction between data reduction and information preservation.

Inventive Principle:
Principle #23Feedback

3Productivity

If captured images are classified into multiple groups based on abnormality degree, then data management efficiency is improved, but system complexity increases

Engineering Contradiction:
Improvedata management efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements segmentation through the classification unit that automatically groups images by abnormality degree. This segmentation improves data management efficiency by enabling targeted storage and retrieval operations, while the automated nature of the classification reduces the operational complexity burden.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The classification unit operates autonomously to organize captured images into appropriate groups based on abnormality detection. This self-service capability improves data management efficiency without requiring external intervention, and the automated classification process manages the complexity internally rather than externally.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4636685A1Inspection system for textile machinery
Publication Date: 2025.10.22 TOYOTA INDUSTRIES CORP
  • EP4636685A1 patent drawingFigure 1
  • EP4636685A1 patent drawingFigure 2~3
  • EP4636685A1 patent drawingFigure 4~5

AI summary

An inspection system (10) inspects a textile machinery (100) based on captured images (P) of the textile machinery (100). The inspection system (10) includes: a classification unit (41) that detects a degree of abnormality in the textile machinery (100) based on the images (P) newly acquired from a camera (11) and classifies each image (P) into an appropriate group of groups based on the degree of abnormality; a storage unit (15) for storing the images (P); a counting unit (42) that counts the cumulative number of the images (P) of a target group and stores the cumulative number of the images (P) of the target group in the storage unit (15); and a determination unit (43) that determines whether the cumulative number of the images (P) of the target group stored in the storage unit (15) has reached an upper limit and determines whether to replace an image of the images (P) stored in the storage unit (15) with the newly acquired image (P).