节点文献
基于NMI特征的ICT断层轮廓识别与处理
Contour Recognition and Processing of ICT Slicing Images Based on NMI Feature
【Author】 HE Shi-han,WEI Dong-bo,GUO Li-ye (School of Mechanical Engineering and Automation,Beihang University,Beijing 100191,China)
【机构】 北京航空航天大学机械工程及其自动化学院;
【摘要】 ICT断层图像的轮廓特征处理技术,在逆向工程、模型三维重建中具有重要的意义,而轮廓整体和分段特征的识别、曲线拟合是其首要的环节。本文提出一种新的综合轮廓处理方法:首先基于NMI(归一化转动惯量)特征,识别出所有闭合轮廓中的圆和正多边形整体特征;再根据曲率差分对剩下的轮廓进行特征点提取和分段特征识别;最后对识别出来的整体和分段特征进行曲线拟合,并调整特征点。实验表明:算法自动化程度高,处理结果准确可靠,精度较高。
【Abstract】 The processing technology of contour feature is important in the reverse modeling and 3D model reconstruction based on ICT(Industrial Computerized Tomography)slicing images.Especially,the whole and segmentation feature recognition and curve fitting are primary part in the process of reverse modeling based on feature.Therefore,a new algorithm for contour processing is proposed.Firstly,the NMI(Normalized Moment of Inertia)feature is used to recognition the closed circle and regular polygon feature.Then,according to the discrete curvature difference,the dominant points recognition and curve type determination of segmented data points are decided.Finally,the curve of the whole and segmentation features is fitted,and the dominant points are adjusted. The experimental results show that the proposed algorithm is effective,accurate and highly automatic.
【Key words】 ICT Slicing Images; Contour Feature Recognition; Normalized Moment of Inertia Feature; Curve Fitting; Curvature Difference;
- 【会议录名称】 全国射线数字成像与CT新技术研讨会论文集
- 【会议名称】全国射线数字成像与CT新技术研讨会
- 【会议时间】2012-07-01
- 【会议地点】中国四川绵阳
- 【分类号】TP391.41
- 【主办单位】中国体视学学会CT理论与应用分会