发动机叶片ICT数字切片图像特征边界的提取

目的 实现数字切片图像特征边界的提取。方法 针对有噪声信号影响的发动机叶片ICT(Industrial computed tomogmphy)数字切片图像，根据图像特征边界点相邻像素灰度值变化特点，通过预先确定提取边界所占矩形区域的大小、方向和预提取图像边界的数量，沿垂直某一矩形边的方向自动产生与提取边界点数量相等的一组等距平行线，并将每一平行线穿过图像的像素灰度变化拟合成一条三次样条曲线或B-样条曲线。结果 根据拉普拉斯边界检测算法在出现边界点处一阶段数为最大和二阶段数为零的特性，计算机边界点的坐标，实现灰度图像特征边界的提取。结论 通过实验得到的测量误差，使用统计的方法进行分析，该算法不但能够达到1／4像素的亚像素精度，而且还有计算效率高的特点。

Detection and extraction of edges in ICT digital images of turbine blades

Aim To detcet the location of edges and extract the coordinates of the characteristics based on the industrial computer tomography(ICT) and its slices pictures with blurring and noise.Methods Within a interactive defined orientable rectangular region of interest,a group of equidistant parallel lines which are vertical to the one side of the rectangular will be created with the line spacing according to the number of finding edge points in this region.Results Every line profile is fitted by using cubic or B-spline and edges are detected by finding zero-crossings,namely the second deriatives of the cubic or B-spline curve are equal to zero.Conclusion The experiment is carried out to prove this approach to be provided not only with one fourth of sub-pixel accuracy for edge detection and extraction,but also with high efficiency and little computationally expense.