中文摘要：

疲劳条带是疲劳断口典型的微观特征,分割是对金属断口图像进行定量分析以反推疲劳寿命和疲劳应力的重要环节.由于断裂过程中的复杂性使得实际断口多表现为多样性的混合形态,且不同区域的疲劳条带周期差别很大,使得疲劳条带纹理区域和纹理边缘的准确定位成为分割的一大难点.传统单一纹理特征对这类复杂的自然纹理分割准确性低.通过分析断口的自然纹理特性,提出结合灰度共生矩阵和小波包变换,采用多特征对断口图像的疲劳条带进行准确分割,从而发挥了时域和频域两类特征的双重优势.实验结果表明,改进的多特征方法对疲劳条带自动分割精度优于传统方法.

英文摘要：

Fatigue striation is the typical microscope characteristic of fatigue fracture, and its segmentation is im portant to the quantitative analysis of metal fracture image, which can backwards concludes the fatigue crack propaga tion life and fatigue stress. Actual fracture is diverse and mixed morphologies, also the period of fatigue striation va ries greatly in different region because of the complexity of fracture process, which make the accurate segmentation of texture region and texture border become a different thing. The accuracy of traditional single texture feature is low for the segmentation of complicated natural texture. According to the analysis on the natural texture characteristics on the fracture surface, this paper develops a method based on multi feature to segment the fatigue striation in metal frac ture image. Gray level co occurrence matrix features and wavelet packet transform features are combined so that the dual advantages of the multi features in time domain and frequency domain are produced. The experimental results show that the method based on multi feature is superior to the traditional methods in the accuracy of auto segmen tation.