基于样本集提取的现场钢材计数研究

ON-SITE STEEL PIPES COUNTING BASED ON SAMPLE SET EXTRACTION

针对现有方法实现现场计数的不可行性,提出一种基于提取样本的方式来实现现场环境下钢管计数方法。首先采用GRA和Top-hat联合增强方式改善图像质量,然后通过采用形态学填充的方式寻找标准物体样本集,最后形成权值模板再对梯度图像进行检测、统计计数。实验结果表明,该方法可以很好地在钢管加工过程及出入库中检测出钢管数量,具有较强的实用性。

Aiming at the infeasibility of current methods in implementing the on-site counting, we propose a sample set extraction-based method to implement the steel pipes counting under the on-the-spot environment. First the method uses GRA and Top-hat combined augmentation means to enhance the image quality, and then by adopting the morphological filling means it finds the sample set of the standard object, finally it forms the weight value template and then tests the gradient image as well as counts the numerical number. Experimental results show that the method successfully achieves the steel pipe counting during the pipe processing and in/out warehouses, and has quite strong practicability.