基于小波定位及Facet模型的三维工业CT图像边缘检测

曾理; 郭海燕; 马睿

doi:null

您当前的位置：

首页 >

文章列表页 >

基于小波定位及Facet模型的三维工业CT图像边缘检测

信息科学 | 更新时间：2020-08-12

- 基于小波定位及Facet模型的三维工业CT图像边缘检测
- Three Dimensional Edge Detection Based on Wavelet Locating and Facet Model
- 光学精密工程 2010年18卷第2期页码：443-450
- 作者机构：
  
  1. 重庆大学ICT研究中心2. 重庆大学ICT研究中心，重庆大学数理学院3. 重庆大学ICT研究中心
  2. 重庆大学数理学院
- 作者简介：
- 基金信息：
- DOI：
  中图分类号：
- 收稿日期：2008-12-23，
  
  修回日期：2009-03-11，
  
  网络出版日期：2010-02-20，
  
  纸质出版日期：2010-02-20
- 稿件说明：
移动端阅览
曾理,郭海燕,马睿. 基于小波定位及Facet模型的三维工业CT图像边缘检测[J]. 光学精密工程, 2010,18(2):443-450

Three Dimensional Edge Detection Based on Wavelet Locating and Facet Model[J]. Optics and precision engineering, 2010, 18(2): 443-450.
曾理,郭海燕,马睿. 基于小波定位及Facet模型的三维工业CT图像边缘检测[J]. 光学精密工程, 2010,18(2):443-450 DOI：

Three Dimensional Edge Detection Based on Wavelet Locating and Facet Model[J]. Optics and precision engineering, 2010, 18(2): 443-450. DOI：

摘要

实际应用中，工业CT三维图像处理对精度和速度都有较高的要求。Facet模型在图像边缘检测方面能够获得精确的边缘信息，但它在进行三维边缘检测时存在非常耗时的缺点，以致很难应用于实际的工业CT三维图像处理。针对这一问题，提出了一种基于3D小波定位的Facet模型边缘检测方法。小波变换边缘检测速度较快，但是得到的边缘依赖于阈值的大小，当阈值较小时边缘较粗，而当阈值较大时边缘容易断裂。本文方法正是利用了该特点，首先对工业CT 三维图像数据进行三维小波变换，设定较小阈值，得到三维粗边缘，即对图像边缘的粗定位，然后再针对粗边缘点逐个进行三维Facet拟合，从而得到实际边缘点，即完成了图像边缘的精确定位。该方法通过小波变换这一前处理过程大大减少了Facet拟合的体素点数，从而使得Facet模型的三维边缘检测得到了加速。实验结果表明，本文方法不仅能得到与直接Facet模型效果相当的边缘，更是大大提高了三维边缘检测的速度，可加速3.51倍至7.39倍

图像边缘越简单加速比越高。这样，使得Facet模型这一精确的边缘检测方法可以应用于实际的工业CT图像三维处理。

Abstract

High accuracy and high speed are both required highly in actual industrial Computed Tomography (CT) three dimensional (3D) applications. Using facet model

we can obtain the precise image edge to meet the needs of high accuracy. But it will consume too much time when be used to detect 3D image edge

so it is difficult to be applied in 3-D edge detection of actual Industrial CT image. Against this problem

we present a method of 3D edge detection based on wavelet translate and facet model. When wavelet translate was be used to detect edge

the speed is high

but the result has relationship with the threshold. If threshold is small

the edge is a little thick that it is hard to locate exactly. If threshold is big

the edge is not continuous that we can’t get the complete contour. However

we just make use of this characteristic to increase the speed of facet model. Firstly

3D wavelet transform is been used to 3D data of Industrial CT

the initial candidates of edge could be obtained by setting smaller threshold. Through this step the edge was been located roughly. Secondly

3D facet model is been used to decide if the candidate is the true edge. Then the 3D image edge of high accuracy is located. The improved algorithm can accelerate the original facet model greatly by the preprocessing of 3D wavelet transform which can reduce the number of edge pixel candidates. Finally

experiments on actual Industrial CT 3D image data and simulated image data validate the efficiency and accuracy of the improved algorithm. The original and improved algorithm were both been realized by programming. Then two groups of experiment result was been obtained. The accuracy of the two algorithm was been compared by computing the area of the same object in the two groups of result. Comparison of the program runtime indicate the speed of improved algorithm is been increased. So the experiment results show that the improved method not only has the equivalent accuracy with direct facet model

but also increases the speed from 3.51 to 7.39 times. The image edge is simpler; the speed of algorithm is faster. In this way

3D facet model which is high accurate can be used into actual Industrial CT 3D image processing. It is of actual significance to Industrial CT high accuracy and speed detection.

关键词

Keywords

references

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

小模数直齿轮免参数并行快速测量

高性能锥束工业X射线CT系统研制与应用

面向小目标测量的通道注意力网络与系统设计

强噪声下自适应Canny算子边缘检测

基于导向滤波Retinex和自适应Canny的图像边缘检测