Assembly Unit Defect Monitoring With Spatial Feature Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing manufacturing processes lack an efficient method to correlate visual and non-visual features of assembly units to identify the origin of defects, leading to inefficient defect analysis and correction.
Innovation Solution
A computer system that leverages optical inspection data to create a feature map linking manufacturing inputs to assembly units in space and time, enabling correlation of visual and non-visual features to isolate the root cause of defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all visual and non-visual features are correlated to identify defect origins, then defect analysis completeness is improved, but computational load and analysis time increase significantly
Solution Approach 1:
The patent segments the feature correlation process by first organizing features into spatial groups based on their location in assembly units, then processing each spatial group separately. This segmentation allows the system to handle large numbers of features more efficiently by breaking down the complex correlation task into smaller, manageable spatial units, thereby reducing overall computational load while maintaining analysis completeness.
Solution Approach 2:
The patent applies local quality by calculating weights for features based on their spatial proximity to hypothesized defect locations. Features closer to the suspected defect origin receive higher weights, while distant features are downweighted or excluded. This localized approach focuses computational resources on the most relevant features, reducing analysis time while maintaining precision in defect origin identification.
2Measurement precision
If all manufacturing inputs are aggregated and correlated, then root cause identification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent extracts and isolates specific spatial groups of features and manufacturing inputs that are most relevant to defect analysis. By filtering and extracting only the features within spatial proximity to hypothesized defect locations, the system reduces the complexity of data processing while maintaining high accuracy in root cause identification. This extraction approach removes unnecessary complexity by focusing only on pertinent data elements.
Solution Approach 2:
The patent introduces a spatial dimension for organizing and weighting features, transitioning from a flat list of all manufacturing inputs to a structured spatial hierarchy. This dimensional change allows the system to process and correlate features based on their physical location in the assembly unit, simplifying the overall system architecture while improving root cause identification accuracy through spatially-aware feature selection.
3Productivity
If computational load is reduced by focusing on proximal features, then processing speed is improved, but potentially relevant distant features may be missed
Solution Approach 1:
The patent applies partial action by selectively processing only the subset of features within spatial proximity to hypothesized defect locations, rather than processing all features. This partial approach focuses computational resources on the most likely candidates, significantly improving processing speed. The system uses spatial weighting to ensure that even within this partial set, all potentially relevant features are adequately considered, balancing speed and completeness.
Data Source
AI summary
One variation of a method for monitoring manufacture of assembly units includes: receiving selection of a target location hypothesized by a user to contain an origin of a defect in assembly units of an assembly type; accessing a feature map linking non-visual manufacturing features to physical locations within the assembly type; for each assembly unit, accessing an inspection image of the assembly unit recorded by an optical inspection station during production of the assembly unit, projecting the target location onto the inspection image, detecting visual features proximal the target location within the inspection image, and aggregating non-visual manufacturing features associated with locations proximal the target location and representing manufacturing inputs into the assembly unit based on the feature map; and calculating correlations between visual and non-visual manufacturing features associated with locations proximal the target location and the defect for the set of assembly units.