涡轮叶片CT图像边缘提取的最优算子研究  

Research on optimal operator of blade edge detection from computed tomography images based on computational theory

作　　者：孙晶晶[1] 杨民[1] 刘静华[1] 

机构地区：[1]北京航空航天大学机械工程及自动化学院,北京100191

出　　处：《航空动力学报》2010年第1期175-179,共5页Journal of Aerospace Power

基　　金：国家自然科学基金(60872080);航天科技创新基金(CASC0410);北京市科技新星基金(2005A14);北京市教育委员会共建项目

摘　　要：为了解决叶片计算机断层(computed tomograghy,简称CT)图像边缘提取精度低效率差的问题,以解析法提取边缘的理论为基础,推导了适用于CT图像边缘提取的最优算子,并通过定量计算与模拟标准图像对最优算子的性能进行了分析,结果表明最优算子在抗噪、定位精度以及单像素响应等方面都比传统的边缘提取算子更为优越.实际应用最优算子提取的叶片边缘更光顺,噪声点以及伪边缘更少,进一步证明了最优算子在CT图像边缘提取方面有着明显的优势.Based on the mathematical model and computational edge detection theory of Canny, an optimal operator for edge detection in computed tomography image processing was proposed in order to resolve the accuracy and efficiency problem in blade edge detection from computed tomography images. The quantitative analysis and simulation of standard images verify that, the performance of the optimal operator is better than the conventional operators in the signal-to-noise ratio（SNR）, location and multi-response. The performance of the optimal operator is also verified by the practical computed tomography images. It is found that the edges of blade detected by the optimal operator are smoother, the noise points and false edges are less.

关 键 词：边缘提取 解析法 计算机断层(CT)图像 最优算子 

分 类 号：V233[航空宇航科学与技术—航空宇航推进理论与工程]