Research On Surface Defects Inspection System Of Steel Strip

The steel strip is one of the main products in the steel industry and as the important raw and processed materials of automobiles, electrical home appliances, shipbuilding, aeronautics and astronautics industry, its surface quality has a direct effect on the product quality. How to inspect these defects during production process on real-time and acquire the information of defects in order to control and improve the quality of steel strip, has been being paid close attention by steel enterprise.The surface defect is always the main limit of improving the quality of steel strip. To inspect the steel strip in real-time manner is an important way to make this better. The defect image and the related information data generated on the product line is very key, especially to the analysis of surface defect and failure of product line.Based on digital pattern process and pattern-recognition the text carry on the surface defect inspect of belted steel. According to the grayscale theoretical foundation, a defect detection algorithm and a method of self-adaptation of the standard drawing cut apart were proposed on the basis of grayscale projection., the grayscale strip image was projected downward and leftward, And then calculate the grayscale mean value of every line and every one, according to maximum difference characteristic of standard grayscale of picture's grayscale projection mean value. if difference exceeds certain range and thinks there was defect, otherwise normal. A large number of experiments prove that this kind of detection method can meet the systematic measuring request well.When find surface defect of belted steel, first cut apart the regional picture of the defect, and then analysed the grayscale value characteristic of the regional picture of the defect, the characteristic of geometry and characteristic of unity and coherence, use the method of the artificial neural network to judge what kind the defect is finally.