压缩感知方法在纺织品有害物质安全检测中的应用

Application of compression sensing method in the safety detection of harmful substances in textiles

摘要:

为提高纺织品生产过程中对有害物质的检测效率，本文将压缩感知方法应用于有害物质的检测过程，以观测矩阵作为混样方案，通过混检的方式得到远少于待测样品的检测次数；混检完成后，根据相应的重构算法由混检值重构出原样品中有害物质的含量，进而获得不合格样品的编号及检出率等；最后，通过仿真实验探讨了由不同参数生成的混检矩阵对重构效果的影响，并用在纤维纺织品中检测有害物质双酚A、芳香胺和甲醛的真实检测项目进行了验证. 验证结果显示, 本文提出的基于压缩感知的混样检测方案不仅能保证检测的准确性，而且能降低检测成本和提高检测效率.

Abstract:

In order to improve the detection efficiency of harmful substances in the textile production process, this paper applies the compression sensing method to the detection process of harmful substances. The observation matrix is used as the sample mixing scheme, and the detection times that are far less than that of the samples to be tested are obtained through the mixed detection. After the mixed detection is completed, the content of harmful substances in the original samples is reconstructed from the mixed detection values according to the corresponding reconstruction algorithm, and then the number and detection rate of unqualified samples are obtained. Finally, the influence of the mixed detection matrices generated by different parameters on the reconstruction effect is explored by simulation experiments, and it is verified by the real detection project of detecting harmful substances bisphenol A, aromatic amines and formaldehyde in fiber textiles. The verification results show that the mixed sample detection scheme based on compression perception proposed in this paper can not only ensure the accuracy of the detection, but also can reduce the detection cost and improve the detection efficiency.