【摘要】 为了实现钢制材料凹坑缺陷的无损检测与形貌反演，基于漏磁检测原理，利用新型单永磁体磁化装置设计了三维漏磁场检测系统，并利用切片实验研究了凹坑缺陷附近的三维漏磁场形貌特征。在漏磁场信号特征分析与提取基础上，利用支持向量机机器学习方法建立了信号特征与缺陷几何参数间的映射模型，进而建立了凹坑缺陷的形貌反演系统。测试实验结果表明，该系统对缺陷长度、深度预测平均误差低于2%,对宽度预测平均误差低于4%,能够较准确地实现凹坑缺陷的形貌反演，展现出广阔的应用前景。更多还原

【Abstract】 To achieve non-destructive testing and morphological reconstruction of pit defects in steel materials, this paper is based on the principle of magnetic flux leakage detection. A three-dimensional magnetic flux leakage detection system was designed using a novel single permanent magnet magnetizing device. Additionally, slice experiments were conducted to study the three-dimensional magnetic flux leakage field morphological characteristics near the pit defect. Based on the analysis and extraction of magnetic flux leakage signal features, a mapping model between signal features and defect geometric parameters was established using support vector machine(SVM) machine learning methods, thereby forming a morphological reconstruction system for pit defects. Test results demonstrate that the system achieves an average prediction error of less than 2% for defect length and less than 4% for width, effectively enabling accurate morphological reconstruction of pit defects and showcasing broad application prospects.更多还原