Container Inspection Image Selection for Rare Defect Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current container treatment plants face challenges in efficiently storing and selecting relevant sensor data for inspection due to the lack of a suitable criterion for image storage, leading to inefficiencies in process monitoring and control.
Innovation Solution
A method for operating a container treatment plant that uses a transport device to capture spatially resolved sensor data, determines a deposit variable based on similarity to reference data, and stores only the most similar data on a non-volatile memory, allowing for improved inspection precision and throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all captured sensor data are stored permanently on memory devices, then complete inspection data is available for analysis, but storage requirements and data management complexity increase significantly
Solution Approach 1:
The patent extracts and stores only the most relevant sensor data by determining a deposit variable for each captured image based on similarity to reference data. This selective extraction approach stores only essential inspection data while discarding redundant information, resolving the contradiction between data completeness and storage requirements
Solution Approach 2:
The patent applies different storage strategies to different data based on their inspection relevance. By calculating similarity variables and deposit variables, the system assigns different storage priorities to different images, storing high-similarity images permanently while managing low-similarity images differently, thus optimizing storage capacity allocation
2Measurement precision
If manual selection and review of captured images is performed, then relevant inspection data can be identified, but the process becomes impractical due to the large volume of images
Solution Approach 1:
The system performs automatic image selection and evaluation through computer-implemented methods that calculate similarity variables and deposit variables. The inspection system self-selects relevant images based on algorithmic criteria rather than requiring manual review, maintaining inspection accuracy while eliminating the time loss associated with manual processing of millions of images
Solution Approach 2:
The patent transforms the manual selection process into an automated parameter-based selection system. By introducing similarity variables and deposit variables as quantitative parameters, the system automatically identifies relevant images based on calculated metrics rather than manual judgment, resolving the time efficiency problem while maintaining precision
3Productivity
If existing image storage strategies based on good/bad rating are used, then basic inspection functionality is maintained, but the system cannot effectively store images with new features or rare defects
Solution Approach 1:
The patent performs preliminary similarity assessment against reference data before final storage decisions are made. By calculating deposit variables based on similarity to known good and bad examples in advance, the system prepares for rare defect detection by pre-identifying images that warrant permanent storage, enhancing both throughput and adaptability to rare defects
Solution Approach 2:
The system uses reference data from previously identified defects and features as feedback to improve future detection. By storing images with high similarity to known rare defects and using this feedback loop, the system continuously improves its ability to detect new and rare defects while maintaining inspection throughput through automated processes
Data Source
AI summary
A method for operating a container treatment plant for treating a plurality of container parts for containers wherein a transport device transports the plurality of container parts as a container part stream along a predetermined transport path from at least one treatment device of a container treatment plant to at least one further treatment device, wherein at least one sensor device for carrying out a container inspection task captures sensor data, and preferably camera images of the container parts. A deposit variable is determined which is characteristic of a deposit instruction for depositing the captured sensor data on a non-volatile memory device. The deposit variable is determined on the basis of a similarity variable which is characteristic of a similarity between the captured sensor data and predetermined and/or predeterminable reference data.