基于图象处理的棒材自动计数技术

AUTOMATIC COUNTING TECHNIQUE FOR STEEL BARS BASED ON IMAGE PROCESSING

利用面阵 CCD摄象机采集打捆后棒材的端面图象 ,通过合适的阈值将灰度图象转换成二值图象 ,以消除棒材图象区域中由于油污造成的孔洞和棒材区域间的粘连现象。根据距离变换思想 ,提出了图象距离概念 ,通过计算并比较棒材区域各象素点的图象距离 ,确定了各根棒材的中心 ,从而实现了从图象中自动计算棒材的数目。对现场采集到的图象进行试验 。

The profile images of bundled steel bars are captured by area scanning CCD cameras, and gray level images are transferred into binary images by adequate threshold, and holes caused by grease on steel bar and joints between bar regions are eliminated. A concept of image distance is introduced based on distance transformation, and image distances of all pixels in steel bar regions are calculated. By comparing these distances, the center of every steel bar is identified. One center means one steel bar, so the number of bundled steel bars can be automatically counted from images. The technique was verified by experiment on a production line of steel bars.