3D Braided Material Trace-Line Measurement With Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual measurement of trace line distances in 3D braided composite materials is cumbersome, time-consuming, and prone to human errors, affecting the quality and reliability of the material, especially in aero-engine applications.
Innovation Solution
A method utilizing a vision data acquisition system with multiple visual sensors and a deep learning model to automatically measure trace line distances, incorporating a high-definition lens, data storage, and computation server for precise positioning and calculation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual measurement method is used to measure trace line distances, then operation simplicity is maintained, but measurement precision and reliability deteriorate due to human errors
Solution Approach 1:
The patent replaces the manual mechanical measurement system with an automated vision-based measurement system. Multiple visual sensors capture images of the 3D braided material surface, and a deep learning model automatically processes these images to identify trace lines and calculate distances, eliminating human error while maintaining operational simplicity through automation.
Solution Approach 2:
The patent creates digital copies of the physical measurement process through vision sensors that capture images of the trace lines. The deep learning model then processes these digital images to extract measurement data, replacing the need for physical contact with the material and enabling non-contact, high-precision measurement.
2Productivity
If manual measurement is performed on complex curved surfaces with corrugated trace lines, then adaptability to complex geometries is maintained, but productivity deteriorates due to time-consuming measurements
Solution Approach 1:
The patent transitions from two-dimensional image capture to three-dimensional surface reconstruction by using multiple visual sensors positioned at different angles. This multi-dimensional approach allows the system to capture the complex curved geometry and corrugated trace lines from multiple perspectives, enabling accurate measurement on irregular surfaces that would be difficult to measure manually.
Solution Approach 2:
The patent performs preliminary actions by pre-training the deep learning model with extensive training data that includes various curved surface geometries and trace line patterns. This pre-training enables the model to quickly and accurately identify trace lines on complex surfaces during actual measurement, eliminating the need for manual adaptation to each new geometry.
3Area of stationary object
If multiple visual sensors are deployed for panoramic coverage, then measurement precision and coverage are improved, but device complexity increases
Solution Approach 1:
The patent merges multiple visual sensors into a unified measurement system that shares common processing infrastructure. The deep learning model processes images from all sensors simultaneously, and the results are integrated to provide comprehensive coverage. This merging approach reduces overall system complexity compared to having separate measurement systems for each sensor.
Solution Approach 2:
The patent creates a universal measurement system where the deep learning model can handle various types of inputs from multiple visual sensors with different fields of view. The same model architecture and processing pipeline are used regardless of which sensors are active or what geometries are being measured, providing multi-functional capability that simplifies system design.
Data Source
AI summary
A method for measuring a distance between trace lines of a 3D braided material, including: (S1) establishing a vision data acquisition system using a vision sensor; (S2) acquiring, by the vision data acquisition system, a training data of the trace lines of the 3D braided material; (S3) constructing a deep learning model for recognizing the trace lines of the 3D braided material; and inputting the training data acquired in step (S2) to the deep learning model to obtain a trained deep learning model; and (S4) positioning a location of the trace lines of the 3D braided material in batch images according to the trained deep learning model obtained in step (S3); and measuring a distance between adjacent trace lines.