摘要
引入核方法分析研究了现有的板坯表面缺陷识别方法,提出了一种新的核函数,并将其应用到板坯表面缺陷特征提取中,用传统的支持向量机对图像进行分类,试验结果表明,新核函数提取的特征识别效果最好,识别率达到了91.55%。
Existed kernel methods were introduced and studied, and a new kernel function was presented and applied to feature extraction of slab surface defects. SVM was used to classify the images. Experimental results show that the feature extracted by new kernel function gets the highest classification rate of 91.55 %.
出处
《物理测试》
CAS
2014年第2期25-27,共3页
Physics Examination and Testing
关键词
板坯
表面检测
支持向量机
核主成分分析
核函数
slab
surface inspection
support vector machine
kernel principal component analysis
kernel function
