Engineering Drawing Validation via Deep Learning Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer systems are unable to automatically validate engineering drawings due to the complexity of the drawings, overlapping annotations, background noise, and structural information loss from Optical Character Recognition (OCR) extraction, which hinders processes like First Article Inspection (FAI).
Innovation Solution
An automatic image validation system that utilizes deep learning models and OCR techniques to extract data from engineering drawings by identifying annotation labels and zone regions, and automatically generates rules for validation based on requirements from related documents, enabling the validation of engineering drawings against specified attributes and specifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If OCR extraction is used to extract data from engineering drawings, then data extraction speed is improved, but extraction accuracy deteriorates due to structural information loss
Solution Approach 1:
The system segments the engineering drawing into multiple zone regions based on annotation labels. Each zone region is processed separately to extract data elements, preserving structural information while enabling efficient extraction. This segmentation approach maintains accuracy by treating different parts of the drawing independently rather than applying OCR to the entire complex drawing at once.
2Measurement precision
If deep learning models are used to identify annotation labels and zone regions, then data extraction accuracy is improved, but system complexity increases
Solution Approach 1:
The system performs preliminary identification of annotation labels and zone regions using deep learning models before data extraction. By pre-identifying these structural elements, the system prepares the drawing data in advance, making the subsequent extraction process more accurate and efficient. This preliminary action reduces the complexity of the main extraction task by organizing the data structure beforehand.
3Productivity
If automatic validation is implemented to validate engineering drawings against requirements, then validation efficiency is improved, but the system requires complex rule generation capabilities
Solution Approach 1:
The system automatically generates validation rules by extracting requirements from related documents and comparing them with data extracted from engineering drawings. The system serves itself by autonomously creating the validation logic without requiring manual rule definition. This self-service approach improves validation efficiency while managing complexity through automated rule generation rather than manual configuration.
4Loss of information
If the system processes overlapping annotations and background noise in engineering drawings, then data extraction completeness is improved, but processing time increases
Solution Approach 1:
The system divides the engineering drawing into multiple zone regions based on annotation labels, processing each region separately. This segmentation allows the system to handle overlapping annotations and background noise more effectively by focusing on specific areas rather than processing the entire drawing uniformly, thus improving data completeness while managing processing time efficiently.
Data Source
AI summary
The validity of engineering drawings is automatically determined based on compliance of the specifications of the engineering drawings with automatically generated rules. A document package including images of the engineering drawings and related documents is received. Rules codifying the requirements to be fulfilled by the engineering drawings are automatically generated from the related documents. Data such as specifications of the various parts of the entities in the engineering drawings are automatically extracted. The extracted data is analyzed to determine compliance with the rules to validate the engineering drawings.