一种基于多光谱图像处理的自适应控制焊缝处理方法
By combining multispectral image processing and adaptive control technology with fuzzy PID control and machine learning algorithms, efficient and accurate automated correction of weld seam inspection is achieved, solving the problems of low efficiency and low accuracy of traditional weld seam inspection, and improving welding quality and connection strength.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ANQING NORMAL UNIV
- Filing Date
- 2024-06-04
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional weld inspection techniques are inefficient and inaccurate, resulting in weld surface defects that affect welding quality and connection strength. Existing methods are insufficient for efficient and accurate defect detection and automated correction.
By employing multispectral image processing combined with fuzzy PID control and adaptive control, weld seam images are acquired through a multispectral camera. LBP feature extraction and SVM algorithm are used to determine weld seam quality. The grinding trajectory is corrected by combining fuzzy logic controller and adaptive control, thereby achieving adaptive planning of the robot's grinding path.
It achieves efficient, accurate, and automated weld inspection, improves the flexibility and robustness of weld quality control, and reduces the impact of external environmental changes on the grinding path.
Smart Images
Figure CN118513935B_ABST