Considering that the surface defects of cold rolled strips are hard to be recognized by human eyes under high-speed cir- cumstances, an automatic recognition technique was discussed. Spectrum images of defects can be got by fast Fourier transform (FFT) and sum of valid pixels (SVP), and its optimized center region, which concentrates nearly all energies, are extracted as an original feature set. Using genetic algorithm to optimize the feature set, an optimized feature set with 51 features can be achieved. Using the optimized feature set as an input vector of neural networks, the recognition effects of LVQ neural networks have been studied. Experiment results show that the new method can get a higher classification rate and can settle the automatic recognition problem of surface defects on cold rolled strips ideally.
An autonatic surface quality inspection system installed on a finishing line of cold rolled strips is introduced. The system is able to detect surface defects on cold rolled strips, such as scratches, coil breaks, rusts, roll imprints, and so on. Multiple CCD area scan canteras were equipped to capture images of strip surface simultaneously. Defects were detected through 'Dark-field illumination' which is generated by LED illuminators. Parallel computation technique and fast processing algorithms were developed for real-time data processing. The application to the production line shows that the system is able to detect defects effectively.
Hao Sun, Ke Xu, and Jinwu XuNational Engineering Research Center of Advanced Rolling, University of Science and Technology Beijing, Beijing 100083, China
