图像分割作为一种基础的分析手段在工业计算机断层(Industrial Computed Tomography, ICT)检测中有着广泛的应用。随着CT扫描方式从二维断层扫描向三维立体扫描的发展, 对于CT图像的分析也由二维图像处理发展为三维序列图像处理, 所以传统的图像分割手段已经不能很好地适应这种需求了。以目前在计算机视觉领域广泛应用的对称区域生长算法为基础, 结合CT图像的特点对参数进行了自动选取, 提出了适用于CT图像分析的对称区域生长分割算法。该方法不依赖于初始种子点的选择, 对所有断层图像进行逐层处理和逐层合并, 有效地降低了内存消耗。试验结果表明, 该方法不仅分割精度高, 而且计算效率也高于传统的分割算法。

Image segmentation has been widely applied in the industrial computed tomography testing(ICT). Along with the ICT　scanning mode developing from two-dimension to three-dimension, the computed tomography image analysis has been changed from two-dimension image to three-dimension image . The traditional image segmentation methods have not kept up with the development yet. A symmetric region growing fast segmentation method is proposed. The parameters used in the algorithm are computed automatically according to the ICT images. And the choice of the initial seed points doesn’t affect the final results. Because this method processes the 3D images and combination the regions slice by slice, the cost of the memory is reduced greatly. The experimental results prove that this method has higher accuracy and computation efficiency than traditional segmentation algorithms.