3D CAD Image Data Generation for Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manufacturing industry faces challenges in generating image data for machine learning due to the difficulty in securing images of on-demand products, especially ships, which have unique shapes and manufacturing processes, limiting the application of machine learning and inhibiting innovation.
Innovation Solution
A method and apparatus for generating image data using 3D CAD data, involving extracting and processing image data from various steps of the manufacturing process, applying actual site information, and creating posture conversion image sets to reflect the manufacturing process, enabling the creation of learning data sets for machine learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If 3D CAD data is used to generate image data for machine learning, then the applicability of machine learning to on-demand products is improved, but the complexity of data processing increases
Solution Approach 1:
The patent segments the complex 3D CAD data processing into multiple distinct steps: extracting first image data from 3D CAD, generating posture conversion image sets, performing image processing through rendering engine, and generating learning data sets. This segmentation makes the complex processing manageable and systematic.
Solution Approach 2:
The patent introduces intermediate structures such as posture conversion image sets and learning data sets as mediators between the raw 3D CAD data and the final machine learning application. These intermediaries simplify the processing by creating standardized formats at each stage.
2Measurement precision
If image data from manufacturing processes is captured, then the monitoring capability is improved, but the security risks increase
Solution Approach 1:
The patent creates virtual copies of manufacturing processes through 3D CAD models and generated image data instead of capturing actual sensitive manufacturing images. This copying approach provides monitoring capability while avoiding security risks associated with capturing real manufacturing process images.
Solution Approach 2:
The patent uses 3D CAD models and rendered images as intermediaries between the actual manufacturing process and the monitoring system. This intermediary layer enables monitoring while preventing direct exposure of sensitive manufacturing information.
3Manufacturing precision
If complete 3D CAD data is processed, then the accuracy of image generation is improved, but the processing time increases
Solution Approach 1:
The patent extracts only the necessary first image data from the complete 3D CAD model that is needed for the current manufacturing step, rather than processing the entire model. This extraction approach maintains accuracy for the relevant portion while reducing overall processing time.
Solution Approach 2:
The patent processes only the partial 3D CAD data corresponding to the current manufacturing step and generates image data for that specific stage, rather than processing the complete model. This partial action provides sufficient accuracy for the current need while significantly reducing processing time.
Data Source
AI summary
Provided are a method and an apparatus for generating image data for machine learning. The method may include extracting first image data for a product generated as a result of conducting a first step among a plurality of steps of a manufacturing process of a product based on 3D CAD data for the product, generating a plurality of posture conversion image sets corresponding to a plurality of postures for the product, respectively, based on the first image data, performing image processing through a rendering engine by applying actual image data to respective images of the plurality of posture conversion image sets, and generating a learning data set for the first image data through the image processing.