基于环状分区和区域生长的混凝土CT图像分割

发布时间：2018-04-09 00:32

  本文选题：混凝土切片　切入点：环状分区　出处：《人民长江》2017年23期

【摘要】：针对混凝土圆形工业CT扫描切片中间暗四周亮、灰度图像无明显双峰的缺陷,为提高图像阈值分割的准确性,设计了一种基于环状分区和区域生长法相结合的图像分割算法,并将其与不同分割方法进行效果对比分析。利用医学软件MIMICS对图像进行选择性编辑、补洞处理和冗余数据去除,并利用布尔操作界定出50μm厚度界面,从而完成混凝土CT切片四值化划分。将四值数据导入ABAQUS中,建立有限元细观模型并进行巴西劈裂试验的数值模拟。数值结果与物理试验结果相近,表明所设计的分割方法能够为真实混凝土精确数值模型的建立提供有效信息。
[Abstract]:In order to improve the accuracy of image threshold segmentation, an image segmentation algorithm based on ring partition and region growth method is designed to improve the accuracy of image threshold segmentation.The results are compared and analyzed with different segmentation methods.The medical software MIMICS is used to selectively edit the image, repair holes and remove redundant data. The 50 渭 m thickness interface is defined by Boolean operation, and the quadrature partition of concrete CT slice is completed.The four-value data were introduced into ABAQUS and the finite element mesoscopic model was established and the Brazilian splitting test was simulated.The numerical results are similar to the physical test results, which show that the proposed segmentation method can provide effective information for the establishment of accurate numerical model of real concrete.
【作者单位】： 西北农林科技大学水利与建筑工程学院;
【基金】：国家自然科学基金资助项目(51379178) 中央高校基本科研业务费专项资金资助项目(ZD2012015)
【分类号】：TP391.41;TU528
，

本文编号：1724076