steel defect detection and classification with combination of sam and yolo v10

Detecting steel surface defects has a key role in providing a high-quality product. in order to control the quality of the steel produced and the final product in the steel production process, it is critical to identify defects in the surface in steel worldwide. in the steel manufacturing industry, producing high-quality steel, spotting surface defects, and separating and recycling the flawed steel components are all very expensive, time-consuming, and challenging tasks. in the proposed method, by combining yolo v10 and sam methods, the test result on the neu-det dataset and severstal dataset shows that the proposed method has better detection accuracy, achieving improvements of 7.5% ap@0.5 and 3.15% ap@50-95 and on the neu-det dataset and 6.16% ap@0.5and 3.05% ap@50-95 on the severstal dataset. our algorithm's usefulness is confirmed by the method's superior performance for steel surface defect identification when compared to the comparison main stream methods.