快速连通域标记算法在堆叠棒材计数中的应用研究

棒材计数问题是棒材生产、销售企业提高自动化水平的重要瓶颈之一。根据棒材堆叠后图像中目标像素的分布形态特征,提出一种基于图段的快速图像连通域标记方法,用以解决工业生产中堆叠棒材的自动计数问题。首先对堆叠棒材截面图像进行二值化、去噪、分割等预处理,尽量减少图像中棒材间的粘连现象,得到较为清晰的二值图像,然后利用快速连通域标记方法对图像进行标记以完成计数。实验结果表明,提出的方法能快速、准确的实现堆叠棒材的自动计数。

Research of Counting Stacked Bars Based on the Fast Connected-component Labeling Algorithm

The problem of bars counting is one of the important bottlenecks for improving automation level for bar producing and selling enterprises. A fast image connected-component labeling method based on runs was proposed to solve the problem of automatic counting of stacked bar in industrial production according to the distribution and morphological characteristics of target pixels in the image after stacking. Firstly, image binaryzation, denoising, segmentation and other preprocessing of stacked bar cross section images were processed so as to minimize the adhesion among bars and get a clear binary image. Then, fast connected-component labeling was applied to label the image to complete counting work. Experimental results show that the proposed method can realize the automatic counting of stacked bar quickly and accurately.