| Popis: |
In this study, low carbon steel (LCS) plates are subjected to cyclic annealing (repeated heating above 910 °C for 30 min followed by furnace cooling), and the variation in microstructure, number of grains and its average size, is determined. The number of grains increases with the annealing cycle, as the grains splits to form newer grains. Similarly, the pearlite concentration, in the low-carbon steel, augments with the annealing cycle. Further, an attempt is made to automatically count the number of grains in the cyclic annealed LCS microstructure by employing image-processing techniques, viz. Canny edge detection and Otsu threshold. The coding for the image-processing algorithms was developed in MATLAB platform. The deviation in the number of grains between the proposed image-processing method and the manual three-circle method is less than 10%. Of the two image-processing methods, Canny edge detection achieves a closer prediction (less than 2%). |
