基于形状先验点阵结构CT图像边缘提取方法  

Edge Extraction Method of CT Image with Lattice Structure Based on Shape Prior

在线阅读下载全文

作  者:高小松 马宁 孙利 刘辉 杨耀东 GAO Xiaosong;MA Ning;SUN Li;LIU Hui;YANG Yaodong(Beijing Spacecrafts Manufacture Factory,Beijing 100190,China)

机构地区:[1]北京卫星制造厂有限公司

出  处:《新技术新工艺》2019年第11期60-65,共6页New Technology & New Process

摘  要:采用增材制造的复杂点阵结构已经得到了广泛的应用,CT检测技术与其他无损检测技术相比,其优势在于能给出与复杂点阵结构的几何结构、组分及密度特性相对应的CT图像。为了准确提取CT图像中点阵结构连杆几何形态特征,提出了一种基于形状先验的CT图像边缘提取方法,利用数学矩的概念将点阵结构感兴趣区等效为具有相同标准二阶中心矩的椭圆,获取点阵结构CT图像感兴趣区边缘各像素点到图像中心的有效数据信息,然后以形状先验信息为检核条件,实现点阵结构CT图像边缘提取。该方法可以不受点阵结构表面不线性、不规则的影响和点阵结构复杂造成的CT图像质量差的影响,准确提取感兴趣区内的CT几何特征,所用参数都是利用数学矩解算,可实现高精度的尺寸测量,并且可实现自动测量。Additive manufacturing lattice structures have been applied wildly,Compared with other NDT techniques,CT detection technology had the advantage of providing CT images corresponding to the geometric structure,components and density characteristics of complex lattice structures.In order to accurately extract the geometric features of lattice structure in CT image,a shape priori based edge extraction method for CT images was proposed.By using the concept of mathematical moments,the ROI of lattice structure was equivalent to an ellipse with second-order central moments.The effective data information from each pixel of the edge of the ROI of dot matrix CT image to the image center was obtained,and then the edge of dot matrix CT image was extracted under the condition of shape priori checking.This method did not be affected by no linear and irregularity of surface of lattice structure including poor quality of CT image of complicated lattice structure,accurately extracted interested CT geometrical characteristic,the used parameters were calculated by mathematics moment,which can realize high-precision dimensional measurement,and automatic measurement.

关 键 词:增材制造 点阵 CT检测 CT图像 形状先验 边缘提取 

分 类 号:TG115.28[金属学及工艺—物理冶金]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象