危险化学品泄漏事故定性侦检

Qualitative detection of hazardouschemicals leak

利用被动式傅里叶变换红外光谱遥测技术,通过建立三层辐射传播模型获取泄漏气体的透过率光谱。将支持向量机应用到气体透过率光谱的定性分析中,通过对训练集PCA降维,获取主成分。利用网格搜索和K折交叉验证获取最优模型参数。通过测试集确定模型的分类效果。结果表明,支持向量机在被动式红外光谱遥测的定性分析中具有很好的应用前景。

Passive Fourier Transform Infrared Remote Sensing Technology （RS-FTIR） was introduced to obtain the transmittance spectra of leaked gas by establishing a three-layer radiation propa gation model. The SVM was applied to the qualitative analysis of the gas transmittance spectrum. The main component was obtained through the training set PCA dimension. The optimal model param eters were obtained by grid search and K fold cross validation. The test set determined the classification effect of the model, The re suits showed that the support vector machine has a good prospect in the qualitative analysis of passive RS-FTIR.