Morphed Surface Defect Removal Using 3D Difference Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for defect removal on manufactured objects with morphed surfaces are inefficient, requiring time-consuming customization and sophisticated knowledge of the manufacturing process, and struggle to distinguish between manufacturing deformations and actual defects.
Innovation Solution
A computer-aided manufacturing system using an image-to-image translation based machine learning algorithm, specifically Conditional Generative Adversarial Networks, to generate a 3D model that removes defects by projecting 2D difference images, differentiating between deformations and defects, and employing modified deconvolution layers to reduce noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing smoothing methods are used to remove defects from manufactured objects, then defects can be removed, but the process requires time-consuming customization and sophisticated knowledge of the manufacturing process
Solution Approach 1:
The patent creates a digital twin or virtual model of the manufactured object that replicates its geometric features and surface characteristics. This virtual model serves as a copy that can be processed computationally to identify and remove defects without physically manipulating the actual object, thereby eliminating time-consuming physical customization steps while maintaining defect removal effectiveness.
Solution Approach 2:
The patent transforms the defect removal problem from physical space to digital parameter space by representing surface geometry as mathematical parameters and deviations. By changing the domain from physical manipulation to computational parameter adjustment, the system achieves automated defect identification and removal without requiring manual customization or specialized manufacturing knowledge.
2Manufacturing precision
If existing smoothing methods are used, then defects can be removed, but they cannot distinguish between manufacturing deformations and actual defects
Solution Approach 1:
The patent segments the surface analysis into distinct components: nominal geometry, manufacturing deformations, and actual defects. By dividing the surface evaluation into these separate elements, the system can identify and process only true defects while preserving intentional deformations, thereby maintaining manufacturing precision without losing deformation information.
Solution Approach 2:
The patent introduces a computational intermediary layer between the raw scanned surface data and the defect removal process. This intermediary analyzes surface deviations, distinguishes between deformations and defects using computational criteria, and selectively processes only actual defects. This mediator prevents loss of deformation information while achieving accurate defect removal.
3Manufacturing precision
If conventional defect removal methods are used, then defects can be removed, but excess material is removed along with defects
Solution Approach 1:
The patent applies local quality analysis by evaluating surface deviations at each specific location independently rather than applying uniform smoothing across the entire surface. This localized approach identifies defects based on their specific geometric characteristics and removes only the minimal material necessary at each defect location, preserving surrounding material and achieving high surface finish quality with minimal material loss.
Data Source
AI summary
Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided repair of physical structures include: generating a two dimensional difference image from a first three dimensional model of at least one actual three dimensional surface of a manufactured object, and a second three dimensional model of at least one source three dimensional surface used as input to a manufacturing process that generated the manufactured object; obtaining from an image-to-image translation based machine learning algorithm, trained using pairs of input images representing deformed and deformed plus surface defected added versions of a nominal three dimensional surface, a translated version of the two dimensional image; generating from the translated version of the two dimensional image a third three dimensional model of at least one morphed three dimensional surface corresponding to the at least one source three dimensional surface. Further, defects can be removed based on the third three dimensional model.