The Research And Realization Of Steel-bar Image Recognition

At present, most of domestic steel factories sale bundling steel-bar, and one of the important links is the counting method, many factories use manual counting, but this method is low efficiency and it can’t keep up with the pace of modern rapid production of steel factory. Therefore, the research on the automatic counting of steel-bar has become an urgent problem, It has important practical significance to improve the counting precision of bundling steel-bar and reduce the labor intensity of workers.This thesis mainly studies using image processing to identify bundling steel-bar image. Firstly preprocessing the bundling steel-bar, convert the color image to gray image, histogram equalization is performed on the gray image, image contrast amplification, and then using the method of combination filter to the image after equalization, it is convenient for further processing; Secondly taking gray image segmentation, this paper use two methods that is the edge segmentation and threshold segmentation. Because the effect of edge segmentation is not ideal, so this paper chooses threshold segmentation method. To the image adhesion problems, this paper used a corrosion-thinning -corrosion mathematical morphology algorithm, eliminates most of the adhesion of target image; Finally using an improved connected region labeling method, according to 8 connected criteria to scan the image, extracting connected region area parameters, using the method of area to complete the recognition and achieve the purpose of counting of bundling steel-bar.This thesis comprehensive pattern recognition, computer vision and digital image processing technology to design the steel-bar counting system based on VC and OpenCV. After a series of image processing to the steel-bar image, finally achieved the purpose of identification and counting function of bundling steel-bar.