Counting Method For Steel Products Based On Digital Image Processing

Usually, we count the amount of steel plates by ourselves. This method is inefficient and inveracious. To solve this problem, a new counting method based on image processing technique is presented in this paper. This method is mainly using hough transformation to extract the line features. Considering image processing can also be used for circle detecting, we do the research on circle detecting based on hough circle transformation ,which can used for detecting of steel bar,steel pipe and even some kind of fruit.The process of line detecting method includes image acquire,image preprocessing,edge detect,hough transform and peak extraction. The profile images of steel plates are captured by CCD camera first. In the image preprocessing the median filtering ,edge detection and image enhancement are conducted. Then the Houghçš„transformation is performed on the contour image to convert the difficult whole detection in image space to easier local peak detecting in parameter space so as to pick up the major straight lines in the image. Contraposed to the problem of"peak clustering", an effective peak extraction method is proposed. And the circle detect method in this paper using the circle hough tranfomation based on gray-scale gradient, whose computation is much more smaller than traditional hough transformation.Finally , the method is programming with matlab, and a lot of experiments have been done based on this. The experimental results show that this counting method is very effective, so it can totally replace of manual count. And the circle detecting method also do well, it is worthy of popularizing and applying.