金属图像中的破损特征分割方法研究  

Research on segmentation method of damage feature in metal image

作　　者：任燕 毕军涛 温雪 

机构地区：青岛酒店管理职业技术学院,山东青岛266100

出　　处：《世界有色金属》2016年第2期94-95,共2页World Nonferrous Metals

摘　　要：采用图像分割方法对金属铸造物进行破损检测，提出一种基于小波尺度分解的金属图像中的破损特征分割方法。首先进行金属图像的纹理分析和识别，采用Harris角点检测方法对图像的破损区域进行角点标定，采用小波尺度分解方法实现破损特征提取和分割。仿真结果表明，该方法进行图像破损区域的特征分割准确度较高。信息提取能力较强。The image segmentation method is used to detect the damage of metal castings, and a new method of image segmentation based on wavelet scale decomposition is proposed. Firstly, the texture analysis and is bound to the metal image, using Harris corner detection method for the image of the damaged area of the corner calibration, using wavelet scale decomposition method to achieve the damage feature extraction and segmentation. The simulation results show that this method is of high accuracy and strong ability of information extraction.

关 键 词：图像分割 小波分析 特征 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]