基于Contourlet和KPCA的焊接缺陷图像特征提取

为了进一步提高焊接缺陷识别的准确度和效率,提出了一种基于Contourlet变换和混沌粒子群优化核主成分分析(kernel principal component analysis,KPCA)的焊接缺陷图像特征提取方法.首先通过Contourlet变换将焊接缺陷图像进行多尺度分解,提取低频分量和特定方向上的高频分量;然后运用混沌粒子群优化后的KPCA分别提取缺陷训练样本和缺陷测试样本的特征;最后根据测试样本特征与训练样本特征之间的欧式距离确定缺陷测试样本的类型.结果表明,与基于核主成分分析特征提取法、基于小波的核主成分分析特征提取法相比,文中方法提取的特征更为完整,识别率更高,运行速度较快.

Feature extraction for welding defect image based on contourlet transform and kernel principal component analysis

In order to further improve the accuracy and efficiency of welding defect recognition, a method was proposed to extract feature of welding defect image based on contourlet transform and kernel principal component analysis(KPCA) by chaotic particle swarm optimization(CPSO). Firstly, multi-scale decomposition of welding defect images was performed by contourlet transform. Low-frequency components and high-frequency components in a certain direction were extracted. Then, features of training samples and testing samples of welding defects were extracted using KPCA by CPSO, respectively. Finally, the type of welding defect testing samples was determined according to the Euclidean distance between features of training samples and features of testing samples. A large number of experimental results show that, compared with the feature extraction method based on KPCA and the feature extraction method based on the combination of wavelet transform and KPCA, the proposed method can extract feature more completely and has higher recognition rate and operating speed.