为提升飞机结构损伤维修分析的智能化水平，实现飞机结构损伤模型的高效重构，提出了基于结构张量特征的动态阈值损伤区域划分方法。首先，引入结构张量理论，特征表示图像局部颜色结构纹理，完备损伤图像特征表示；其次，通过构建共性分布的结构张量特征空间，从而使不同损伤图像具有一致性损伤区域划分流程；在此基础上，定义了动态阈值划分算子，并通过对动态阈值划分算子参数的计算，实现结构张量特征空间中的动态阈值划分；最后，选用不同飞机结构损伤实例图像对方法进行了对比验证。验证结果表明，相比于传统灰度方法、固定阈值划分和其他动态阈值划分算子，椭圆算子划分得到的损伤边界连贯完整，能够有效分割微小裂纹，噪点较少，质量较优。该方法运算高效、流程统一、特征表示完备，对不同类型结构损伤的适用性好。
To improve the intelligentization level of maintenance analysis of aircraft structural damage, and realize the reconfiguration of aircraft structure damage model efficiently, an airframe damage region dynamic threshold division method based on structural tensor characteristics is proposed. By introducing the structural tensor theory, local image color structure and texture features are represented to complete the feature representation of damage images. And by constructing the same-distribution structure tensor feature space, the damage region division consistency process of different damage images is constructed. On the basis, the dynamic threshold division operators are defined. By computing the parameters of dynamic threshold division operators, the dynamic threshold division of structural tensor feature space is realized. Finally, a contrast verification is performed using different airframe structural damage images. The result shows that comparing with the traditional gray method, the fixed threshold division method and the other dynamic threshold division operators, ellipse operator shows better quality of division. And its damage region boundary is coherent and complete. Tiny cracks can be divided effectively with less noise. The method efficient in operation, and consistent in process showing high applicability to different types of structural damage.
[1] SHENG H, DENG S, ZHANG S, et al. Segmentation of light field image with the structure tensor[C]//IEEE International Conference on Image Processing. Piscataway, NJ: IEEE Press, 2016: 1469-1473.
[2] CONDESSA F, BIOUCAS-DIAS J, KOVACEVIC J. SegSALSA-STR: A convex formulation to supervised hyperspectral image segmentation using hidden fields and structure tensor regularization[J]. Computer Science, 2015, 19(12): 2063-2075.
[3] 谢晓振, 吴纪桃. 基于结构张量和Wasserstein距离的纹理图像分割[J]. 中国图象图形学报, 2010, 15(9): 1345-1351. XIE X Z, WU J T. Texture segmentation based on the use of structure tensor and the Wasserstein distance[J]. Journal of Image and Graphics, 2010, 15(9): 1345-1351 (in Chinese).
[4] GE Q, XIAO L, ZHANG J, et al. An improved region-based model with local statistical features for image segmentation[J]. Pattern Recognition, 2012, 45(4): 1578-1590.
[5] 张善卿, 张坤龙. 基于结构张量特征值的纹理图像分割模型[J]. 电子学报, 2013, 41(7): 1324-1328. ZHANG S Q, ZHANG K L. Texture image segmentation model based on eigenvalues of structure[J]. Acta Electronica Sinica, 2013, 41(7): 1324-1328 (in Chinese).
[6] YIN X M, WEI M, YAO Y H, et al. A variational framework for multi-region image segmentation based on image structure tensor[C]//Chinese Conference on Image and Graphics Technologies. Berlin: Springer, 2013: 260-268.
[7] YUAN J, WANG D, CHERIYADAT A M. Factorization-based texture segmentation[J]. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2015, 24(11): 3488-97.
[8] MEWADA H, PATEL R, PATNAIK S. A novel structure tensor modulated Chan-Vese model for texture image segmentation[J]. Computer Journal, 2015, 58(9): 1-17.
[9] 李梦. 图像分割的结构张量几何活动轮廓模型[J]. 计算机应用研究, 2014, 31(12): 3890-3893. LI M. Geometric active contours based on structure tensor for image segmentation[J]. Application Research of Computer, 2014, 31(12): 3890-3893 (in Chinese).
[10] HAN S, XU W, TAO W, et al. Color-texture cosegmentation based on nonlinear compact multi-scale structure tensor and TV-flow[J]. Signal Processing, 2016, 131(1): 456-471.
[11] ZHANG Y, YUAN J, LIU H, et al. GrabCut image segmentation algorithm based on structure tensor[J]. The Journal of China Universities of Posts and Telecommunications, 2017, 24(2): 38-47.
[12] XU W, HAN S D, PENG Y C. Texture segmentation based on nonlinear compact multi-scale structure tensor and TV-flow[C]//Eighth International Conference on Digital Image Processing. Pittsfiela, MA: International Society for Optics and Photonics, 2016: 100331A.
[13] 杨勇, 郭玲, 王天江. 基于多尺度结构张量的多类无监督彩色纹理图像分割方法[J]. 计算机辅助设计与图形学学报, 2014, 26(5): 812-825. YANG Y, GUO L, WANG T J. Multi-scale structure tensor based unsupervised color-texture image segmentation approach in multiclass[J]. Journal of Computer-Aided Design and Computer Graphics, 2014, 26(5): 812-825 (in Chinese).
[14] ZHANG W, FEHRENBACH J, DESMAISON A, et al. Structure tensor based analysis of cells and nuclei organization in tissues[J]. IEEE Transactions on Medical Imaging, 2016, 35(1): 294-295.
[15] WANG X, FANG S, ZHU X, et al. Nonlinear diffusion and structure tensor based segmentation of valid measurement region from interference fringe patterns on gear systems[J]. Current Optics & Photonics, 2017, 6(1): 587-597.
[16] NERGIZ M, AKIN M. Retinal vessel segmentation via structure tensor coloring and anisotropy enhancement[J]. Symmetry, 2017, 9(11): 276-278.
[17] 蔡舒妤, 师利中. 一种改进谱聚类的机体损伤图像过渡区提取方法[J]. 计算机辅助设计与图形学学报, 2016, 28(10): 1732-1739. CAI S Y, SHI L Z. An airframe damage image transition region extraction method based on improved spectral clustering[J]. Journal of Computer-Aided Design & Computer Graphics, 2016, 28(10): 1732-1739 (in Chinese).
[18] OSMAN M Z, MAAROF M A, ROHANI M F. Improved dynamic threshold method for skin colour detection using multi-colour space[J]. American Journal of Applied Sciences, 2016, 13(2): 135-144.
[19] KUMAR G, SHARMA S. A localized region based active contour method for image segmentation using dynamic threshold[J]. International Journal of Computer Applications, 2017, 166(10): 31-35.
[20] ZHANG S Q, WANG W L, LU J F, et al. Image segmentation using a new scalar texture descriptor based on extended structure tensor[J]. Journal of Information Hiding and Multimedia Signal Processing, 2016, 7(5): 1020-1030.
CJA