基于断口图像纹理特征提取仿真方法研究  

Research of Fracture Image Texture Feature Extraction Based on Wavelet Analysis

作　　者：何其贵[1] 余春平[1] 余丽萍[1] 

机构地区：[1]江西信息应用职业技术学院,江西南昌330043

出　　处：《计算机仿真》2011年第9期279-282,共4页Computer Simulation

摘　　要：研究准确识别断口图像特征很重要,断口图像的识别是进行金属材料失效分析的重要组成部分,而图像纹理特征的提取则是识别的最关键的环节。应对金属断口的特征、性状、形貌等进行识别。为了提高识别断裂机理和准确识别断口形态,可而更好地指导生产实践,并为机械产品的设计和制造提供依据。提出采用小波分析方法,对金属断口图像进行分解、提取其能量纹理特征进行了相关实验,并在分类中进行实验,证明了小波分析提取的纹理特征对金属断口图像分类精度和识别精度都得到了提高。Fracture Image recognition of metallic materials is an important part of failure analysis, and image tex- ture feature extraction is the key of the identification. The metal fracture characteristics, traits, morphology and other issues should be recognized. In order to better guide the production practice, the mechanical product design and man- ufacturing provide the basis. Based on wavelet analysis, the article decomposes the image, and extracts the energy texture features to implement related experimental analysis, and its application in the classification experiments shows that the classification accuracy of texture features of metal images based on wavelet analysis is improved.

关 键 词：小波分析 纹理特征 提取 研究 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]