中文摘要：

为快速、安全地检测氧乐果和毒死蜱农药残留，使用改进的氯化钯比色法，分别采集2种农药比色反应后的吸收光谱，利用主成分分析法（PCA）和偏最小二乘法（PLS）建立预测模型。结果表明：1）使用乙酸代替传统浓盐酸配制的氯化钯比色试剂效果更理想，氧乐果和毒死蜱的吸收光谱可以区分的检测下限分别是0．05和0．50mg／L，基本满足GB2763—2014对部分果蔬的检测要求。2）氧乐果和毒死蜱吸收光谱的敏感波长分别为510和499nm，由此确定最优建模波段。对比PCA和PLS方法建模效果，2种农药的原始光谱数据都是在Savitzky—Golay（SG）平滑处理后，使用PLS方法在主成分为4时预测效果较好。PLS模型在2-折交叉验证下，氧乐果在480～680nm波段，建模集相关系数Rc=0．9770，均方根误差RMSEC=5．801，验证集相关系数Rp=0．9630，均方根误差RMSEP=7．904；毒死蜱在460～850nm波段，Rc=0．9970，RMSEC=2．281，Rp=0．9847，RMSEP=3．170。该方法比色试剂配制简单安全，室温下比色反应时间为2min，满足快速、安全的检测要求，为进一步研究实用的农残检测仪器提供了新方法。

英文摘要：

In order to rapidly and safely detect omethoate and chlorpyrifos pesticide residues, using improved PdCl2 colorimetry,prediction models of the two pesticides by principal component analysis （PCA） and partial least squares （PLS） were established respectively with absorbance spectra after corresponding coloring reaction. The results showed that：1） The effect of the color agent PdCI2 which was dissolved in acetic acid instead of the common concentrated hydrochloric acid was more ideal. For omethoate samples, the mass concentration of 0.05 mg/L was the minimum value that could be distinguished in the spectra and for chlorpyrifos samples and the minimum mass concentration was 0.50 mg/L. The results basically met the pesticide residue detection requirements of some fruits and vegetables according to China National Standard GB 2763-2014; 2） Sensitive wavelength of omethoate and chlorpyrifos absorbance spectra were 510 and 499 nm respectively,which determined the optimal modeling waveband. The effect of prediction model by PLS was better than by PCA where the number of principal components was four after Savitzky- Golay （SG） smoothing for the two pesticides. By 2-fold cross validation,the optimal prediction model of omethoate was obtained in the region of 480- 680 nm, the correlation coefficient of calibration set Rc = 0. 977 0, root mean squared error of calibration set RMSEC = 5. 801; the correlation coefficient of prediction set Rp = 0. 963 0, root mean squared error of prediction set RMSEP = 7. 904. The optimal prediction model of chlorpyrifos was obtained in the region of 460 - 850 nm,Rc= 0.997 0,RMSEC = 2. 281; Rp = 0. 984 7,RMSEP= 3. 170. The results determined that this method was safe in operation and the colorimetric reaction time was 2 min at room temperature, which met the requirement of fast and safe detection and provided a new method for further development of practical pesticide residues detection instrument.