论文部分内容阅读
分析了支持向量机的分类原理,指出在各类别样本数目相差较悬殊时,SVM不能获得良好的分类能力。针对焊接缺陷分类,提出了加权SVM(WSVM)算法。测试结果表明,该算法在焊缝RT图像中缺陷的分类识别中,能提高小类别缺陷的的检测精度,具有较高的理论和应用价值。
The classification principle of SVM is analyzed, and it is pointed out that SVM can not obtain good classification ability when the number of sample in each category is more than one. According to the classification of welding defects, a weighted SVM (WSVM) algorithm is proposed. The test results show that the proposed algorithm can improve the detection accuracy of small class defects in the classification and identification of defects in weld seam RT images, which has a high theoretical and practical value.