Hydraulic metal structure corrosion detection and sample expansion method and system
By using the dual Transformer network method in the corrosion detection of hydraulic metal structures, the problems of low detection efficiency and low accuracy in the prior art are solved, and efficient and intelligent corrosion detection and sample expansion are achieved.
CN119672430AActive Publication Date: 2025-03-21DALIAN UNIV OF TECH
5 Cites 0 Cited by
Patent Information
- Application Number
- CN202411768121.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-04
- Publication Date
- 2025-03-21
- Estimated Expiration
- 2044-12-04
AI Technical Summary
Technical Problem
The prior art has low detection efficiency and low accuracy in hydraulic metal structure corrosion detection, and manual detection has strong subjectivity.
Method used
The hydraulic metal structure corrosion detection and sample expansion method based on the dual Transformer network is adopted. Through data collection and labeling, data preprocessing, Transformer network optimization and training, automatic segmentation of corrosion areas and intelligent determination of corrosion degree are achieved.
Benefits of technology
It improves the efficiency and accuracy of corrosion detection, reduces the subjectivity of manual detection, and realizes automatic identification of abnormal detection results and semi-automatic expansion of difficult samples.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses a hydraulic metal structure corrosion detection and sample expansion method and system. The method comprises the following steps: S1, data collection and labeling; s2, preprocessing the data set; s3, optimization and improvement of the Transform network are carried out; s4, carrying out Transform network training; s5, rust detection; s6, judging a corrosion detection result; and S7, processing a detection result. According to the method, the segmentation network based on the Transform architecture is used for segmentation and identification of the corrosion area, and the classification network based on the Transform architecture is used for judgment of the corrosion degree. According to the method and the system, two tasks of corrosion area segmentation and corrosion degree judgment can be realized in parallel, compared with an existing algorithm, the corrosion detection precision is improved, and more corrosion detection information can be obtained. And the automatic expansion of difficult rusting samples in the detection process can be realized, and the method is used for network iterative learning.
Need to check novelty before this filing date? Find Prior Art