The Applied Research Of Bar Automatic Counting's On-line Image Processing Methods

The quantificational packing of reinforcing steel bars in steelworks is still staying at a low level. Reinforcing steel bars are counted by labor. This work is so monotony that mistakes always take place and leave trouble to the following business of selling. So developing an automatic system for reinforcing steel bar counting shall meets practical need of steelworks. For the plant's situation is very complicated and special, the attempt of reinforcing steel bar counting based on neither photoelectric sensor nor weight sensor can act on. In this paper, a new method of auto-counting based on digital image processing is brought forward. This method, which counts reinforce steel bars on transmission belt before packing, quite different from that after packing, is significative for the automatization of reinforcing steel bar product line and convenience for producing and selling.In order to achieve this goal of steel bars'automatic counting, diagnostic engineering center on equipment in Northeastern University and FuShun steel Co. Ltd cooperated on the automatic counting system online for steel bars. Base on the purpose of automatically counting out the number of a bundle of Steel-bar by means of digital image process, this paper is mainly research the relevant image process methods. The main contents of this paper contains three parts, first part introduce how to build up such Image Sample Dispose System and the flow of its work step; then give the improved methods of segmentation and morphology in such system; discuss the research of target ladling and counting methods.The emphases of this paper are focused on the algorithms of image processing. The images of reinforcing steel bars are divided into two types. One type of images, in which steel bars'cross sections are brightness, are segmented by gray thresholding value, and the other type of images, in which reinforcing steel bars'cross sections are colorful, are segmented by Fisher algorithms. The segmentation results are all binary images.This article serves the counting purpose through the recognition of the steel bars'image, its main research content including image pretreatment, image division and pattern recognition. In the part of image pretreatment, this paper focuses on the binary image's transform algorithm. In the part of image segmentation, this paper is based on the edge detection algorithm such as Roberts operator, Sobel operator, Prewitt operator, Krisch operator, Canny operator and so on. The Canny operator is the most important of all in this paper. In the image recognition counting part, In this paper, the use of the seed law and the search area of law, with a focus on seed law of search, firstly, doing distance transformation to the processed binary images, then you can get the cores of all regions, search on these core spots again and achieve the bar count at last.