节点文献

基于工业CT的管道圆柱度测量与内表面显示算法研究

Study on Cylindricity Error Evaluation and Inner Surface Display of Pipe Based on Industrial CT

【作者】 刘勇

【导师】 曾理;

【作者基本信息】 重庆大学 , 计算数学, 2010, 硕士

【摘要】 工业CT,即工业计算机层析成像(Industrial Computerized Tomography),是指在工业中用射线扫描待测物体,获得物体断层的投影数据,经重建后得到一系列能够反映待测物体内部结构的切片图像。基于工业CT的检测技术有着无破坏性的显著优点,能够解决许多传统方法无法解决的检测难题。工件中常包含各种各样的管道,对管道的圆柱度误差测量和内表面显示对评定工件的整体性能和使用寿命有着重要意义。本文以工业CT为基础,研究了工件内部管道的圆柱度测量与内表面显示算法。圆柱度误差是指实际被测圆柱面对其理想圆柱的变动量。由于测量工具的限制,传统方法对无法充分接触的内部管道的圆柱度难以测量。本文研究了一种借助工业CT进行圆柱度测量的方法,首先对待检测工件进行CT扫描,得到工件的CT图像,然后提取待测目标的边缘点,采用最小二乘圆柱法建立数学模型,通过优化求解得到圆柱度误差。实验结果表明,该方法可以在无接触无破坏的前提下得到较为准确的测量结果。对工业CT图像进行分析和处理的过程中,经常需要对工件管道的内表面进行显示,为后续的分析与检测做准备。研究了内表面显示的两种方法:剖开显示与虚拟内窥显示。针对剖开显示,首先采用区域生长法分割目标,经轮廓提取和轮廓跟踪定位目标边缘点,然后计算各边缘点到形心的距离。将此距离经过线性变换转换为灰度信息,写入图像。实验结果表明,剖开显示简单有效,可以观测到内表面的缺陷。虚拟内窥显示能够模拟传统光学内窥镜更加立体地显示目标内表面。首先对目标对象进行分割,同时定位目标边缘点坐标。然后提取目标的中心轴径,为了使虚拟镜头能够沿着中心轴平稳前进,用B样条插值对中心轴径作平滑处理。在已知目标边缘点坐标的基础上,采用最短对角线法完成轮廓拼接并最终实现绘制。针对管道的柱状结构,研究了一种特定的轮廓线拼接方法。虚拟内窥显示可动态地观测到管道内表面,更为形象直接。

【Abstract】 Industrial computerized tomography (Industrial CT) acquires projection data, which the x-ray transmits the cross-section of the tested object without damage to the object. The cross-section images can be obtained with special image reconstruction algorithm. Industrial CT, which has the obvious advantage of non-destructivity, can solve many problems that the traditional methods can not solve. There are many kinds of pipelines in real workpiece. The cylindricity error evaluation and inner surface display of pipe are significant for evaluating performance and service life. Based on industrial CT, this dissertation studied the cylindricity error evaluation and inner surface display of pipeline.Cylindricity error is defined as the changes between measured column and ideal column. Because of the limit of measuring tools, the traditional methods can not evaluate the cylindricity error of internal pipe. This dissertation presents cylindricity error measure according to industrial CT. The inspected area is scanned by industrial CT, which obtains the CT images of the inspected area. The edge is detected with Facet algorithm. Then the mathematical model is established by the least-square method. The cylindricity error can be got by optimization. The results show that this method can get cylindricity error without damage and touch.In industrial CT image analysis and process, it is necessary to display the inner surface of the pipeline for subsequent analysis and defect detection. There are usually section display and virtual endoscope for inner surface display. At first, we get the target with region growing for section display, then use contour extracting and contour tracking to get the target’s contour points. Finally, we calculate the distances between the centroid and the contour points and transform the distances to gray for writing to an image. The experimental results show that section display is simple and effective. It can observe the defects on the inner surface.Virtual endoscope display is a kind technology of simulating traditional optical endoscope. It can display inner face of target with three-dimensional quality. At first we use image segmentation to get the target and locate the edge points, then extract the central path of the target. In order to move the virtual camera lens slowly we smooth the path with B-spline. After that we use the shortest diagonal method to complete the contour tiling and render the surface at last. Also we study a special method for the contour tiling of the pipeline. Virtual endoscope display can observe the inner surface of the pipe dynamically and directly.

  • 【网络出版投稿人】 重庆大学
  • 【网络出版年期】2011年 03期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络