Research on an Automatic Counting Method for Steel Bars' Image
Research on an Automatic Counting Method for Steel Bars' Image
复制标题
DOI：
10.1109/icece.2010.405
复制
发表时间：
2010-06
期刊：
影响因子：
--
通讯作者：
Xin Ying;Xue Wei;Yuan Pei-xin;Han Qing-da;Chen Chang-hai
中科院分区：
文献类型：
--
作者：
Xin Ying;Xue Wei;Yuan Pei-xin;Han Qing-da;Chen Chang-hai
针对目前钢厂棒材计数问题，提出了一种棒材图像的自动分割方法。用链码法测量钢筋的近似半径。利用Sobel算子检测钢筋的边缘，利用大津法进行钢筋与背景的分离。然后采用改进的Hough变换对钢筋图像进行灰度累加，通过阈值分割得到钢筋的位置，最后利用最大似然法对钢筋图像进行分割。找到区域中的灰度值。经过实验，实现了钢筋等近似圆图像的精确计数方法。
A method for automatic segmentation of steel bars' image is put forward to solve the counting problems of steel bars in the steel factories at present. The approximate radius of steel bar is measured by chain coding. The edges of steel bars are detected by Sobel operator, and separation of steel bars and background is carried out by Otsu method. Then the improved Hough transformation is adopted in the image of steel bars in order to accumulate the gray values, and the locations of steel bars are got by the threshold segmentation, then max. gray values in a region are founded. After experiments, the accurate counting method of approximate circle images such as steel bars is realized.