Fluorescent Magnetic Particle Inspection Defect Recognition System Image Processing Technology

Magnetic crack inspection is a kind of general Non Destructive Testing (NDT) methods. It is used to detect the surface cracks of ferromagnetic material. Fluorescent magnetic particle inspection has been widely used because of its high sensitivity and its brief and dependable craft. The studying of automatic inspection system is the main difficulty for fluorescent magnetic crack inspection.In this paper, the characteristics railway wheelsets images are analyzed, the characters of the defects on railway wheelsets surface are investigated, and the defects are sorted. Digital image segmentation methods for fluorescent magnetic particle indications are designed, including efficient thresholding segmentation methods and edge-based segmentation methods. Crack inspection methods based on improved morphological grads and multi-blocks extremums are presented and realized based on logical hypotheses to railway wheelsets images. On the base of analyzing the unique characteristic of railway wheelsets and its surface cracks, feature extraction and crack identification methods of segmented images are presented. Synthesizing the characteristic of inspection pipelining and the research all above, fluorescent magnetic crack inspection system is designed. The system plans the pipelining flow reasonable and makes the system models work in phase. At last, the fluorescent magnetic crack inspection system is realized and runs on the inspection workplace successfully.Studies in this paper are a good start for the development of automatic fluorescent magnetic crack inspection system. Meanwhile, the system designed in this paper could be used in real inspection pipelining.