A method for detecting defects in a weld of a pressure vessel based on machine vision
By projecting grating stripe images and constructing a data reliability weight map and a topology skeleton map, the problem of detection deviation caused by uneven illumination and complex texture in weld inspection is solved, and higher accuracy defect identification is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 山东中福环保设备有限公司
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-17
AI Technical Summary
Existing industrial vision technologies struggle to form a stable spatial structure perception in weld inspection under conditions of uneven lighting, reflection interference, or complex textures, leading to deviations and misjudgments in inspection results.
By projecting grating stripe images onto the surface of the pressure vessel weld, collecting pixel brightness values, generating a truncated phase space matrix, constructing a data reliability weight map, obtaining suspected defect morphology maps, stripping outer pixels while retaining central connected branch lines, generating a defect topology skeleton map, calculating the topology entanglement structure index, and outputting defect control commands.
It improves the anti-interference ability, spatial expression ability and structural discrimination ability of weld defect detection, and improves the accuracy of detection results.
Smart Images
Figure CN122415461A_ABST