Image-Based Adhesive Selection for Fractured Part Assembly
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to accurately determine the best adhesive for joining sub-objects of a broken object based on its fractures or cracks, and do not provide optimal glue usage or assembly instructions.
Innovation Solution
A method using image processing and machine learning to analyze the surface porosity and characteristics of sub-objects, determining the suitable adhesive and quantity needed, and providing assembly instructions through augmented reality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing is used to analyze surface porosity and determine adhesive selection, then adhesive selection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent replaces manual adhesive selection with an automated image processing system that captures images of the broken object, analyzes surface porosity characteristics, and uses machine learning models to recommend appropriate adhesives. This substitution of mechanical/manual processes with automated optical and computational systems resolves the contradiction by improving measurement precision through objective image analysis while managing device complexity through software-based solutions.
Solution Approach 2:
The system enables self-service adhesive selection by allowing users to simply capture an image of the broken object with their smartphone. The machine learning model automatically analyzes the image to determine surface porosity and recommends suitable adhesives without requiring users to manually assess surface characteristics or search through adhesive options, thereby improving accuracy while keeping the user-side device complexity minimal.
2Measurement precision
If multiple digital images are captured from different angles and lighting conditions, then surface feature identification accuracy is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary actions by automatically processing and analyzing multiple images captured from different angles and lighting conditions through machine learning algorithms. This preliminary automated analysis extracts surface porosity characteristics efficiently, improving identification accuracy while minimizing the time burden on users since the complex multi-image processing is handled automatically by the system rather than requiring manual review of each image.
Solution Approach 2:
The system creates digital copies of the object's surface from multiple image perspectives and uses machine learning to synthesize this information. By working with digital copies rather than requiring physical inspection from multiple angles, the system improves surface feature identification accuracy while significantly reducing the time required compared to manual multi-angle inspection.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to a method for image processing is described, the method comprising determining an object with a plurality of sub-objects to be joined in each digital image of a plurality of digital images containing the object, determining one or more pore sizes of surface materials of the sub-objects to be joined of the determined object, and determining, using the one or more pore sizes and by means of a model, an adhesive to be used for joining the sub-objects from a set of a plurality of predetermined and stored adhesives.