基于视觉识别的产品平整度质量检测处理系统及方法
By combining a visual recognition system with convolutional neural networks and fractal calculations, the problem of low detection accuracy for complex surfaces has been solved, enabling high-precision flatness detection of steel plate surfaces and ensuring product quality.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING HEXIN AUTOMATION CO LTD
- Filing Date
- 2025-06-06
- Publication Date
- 2026-07-17
AI Technical Summary
Existing inspection technologies based on structured light and triangulation cannot accurately detect the flatness of steel plate surfaces when faced with complex defects, especially in the presence of oil stains, rust, edge laser cutting slag, and surface scratches, which leads to a decrease in inspection accuracy and fails to meet the requirements of high precision and high automation.
A product flatness quality inspection system based on vision recognition is adopted. Combining convolutional neural network algorithm and fractal dimension calculation, image data is acquired through a structured light camera to identify patch features and defects on the workpiece surface. Numerical interpolation of defect areas is performed through an LSTM model. Combined with fractal calculation and reflectivity correction, the inspection accuracy is improved.
It significantly improves the detection accuracy of complex surfaces, reduces errors caused by vibration and uneven speed, accurately identifies complex patch features such as oil stains and rust, improves the accuracy and precision of detection, and ensures product quality.
Smart Images
Figure CN120740502B_ABST