一种钛合金微观结构图像的组织特征量化提取方法及系统
By constructing the TiGrainsU-Net model and combining it with ResNet and Canny edge detection algorithms, the problem of insufficient detailed segmentation capability of titanium alloy microstructure image segmentation models was solved, and high-precision quantitative extraction of tissue features was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2025-06-30
- Publication Date
- 2026-07-17
AI Technical Summary
Existing image segmentation models suffer from insufficient detailed segmentation capability and accuracy in titanium alloy microstructure images, making it difficult to effectively extract quantification parameters of microstructure features.
We employ a deep learning segmentation model U-Net combined with an attention mechanism and a traditional boundary enhancement algorithm. By constructing the TiGrainsU-Net model and combining it with ResNet for grain image classification, we introduce multi-level denoising and Canny edge detection algorithms to optimize the segmentation results.
It improves the grain boundary recognition accuracy and segmentation accuracy of titanium alloy microstructure images, realizes the automation and high-precision quantification of microstructure features, enhances the ability to extract boundary features, and reduces misjudgments.
Smart Images
Figure CN120823598B_ABST