图像阈值分割方法运用于中空纤维膜截面尺寸的测量

Measurement of Cross-Sectional Dimension for Hollow Fiber Membrane by Image Threshold Segmentation Method

比较了迭代阈值分割法、Otsu阈值分割法和最大熵阈值分割法对中空纤维膜截面尺寸测量的影响。试验证明,三种阈值分割方法均能对膜截面图像进行有效分割并进而完成准确测量。最大熵阈值分割方法相比较于目视测量参考值,外径和内径的平均相对误差分别为0.82%和1.41%,更适合膜截面尺寸的准确测量。

Influence of cross-sectional dimension measurement for hollow fiber menlbrane by different threshold segmentation methods is studied, including iterative thresholding, Otsu＇s thresholding and maximum entropy thresholding. Three threshold segmentation methods can effectively segment cross-sectional dimension image of hollow fiber membranes, then cross-sectional dimension measurement is conducted. As maxinmm entropy thresholding method is compared with manual reference value, relative error of external and inner diameters are only 0. 82% and 1.41% respectively. It indicates that maximum entropy thresholding is more suitable for accurate measurement of cross-sectional dimension.