Research of Counting Method of Bars Based on Concave Dots Searching

In view of problems that current counting methods of bundles of bars have low rate of calculation, great color bias and large calculated amount, and are easy to be influenced by vague or incompletion edge, the paper proposed a new counting method of bars by dividing adherent bars and counting connected areas. Firstly, the method pre-processes images of bundles of bars and uses Canny edge detection to obtain edge contour of adherent bars, then scans concave areas in the contour, takes the close concave dots as a concave dot group, and takes the most prominent point as a concave dot. Next, it uses concave dot matching condition to segment the image with those successfully matched concave dots. Finally, it counts connected areas segmented to achieve counting purpose. The experiment result showed that the method has a high accuracy for segmentation of adherent bars, and the segmented edge of image is very similar to the real edge of bundles of bars. The accuracy ratio of segmentation for steel bars whose diameters are above 12 mm is than 99.99%, while the one for steel bars whose diameters are between 10 mm and 12 mm is more than 99.80% without blooking, which satisfies the accepted standard of enterprises.