利用在600～1 100 nm波段范围内可见-近红外反射光谱分析技术, 对常见的高残留农药在绿色植物活体上的无损检测进行了研究。 首先将采集到的漫反射光谱数据进行小波变换提取光谱特征, 然后再利用主成分分析方法进一步对光谱特征进行分析, 最后把这些光谱的前两个主成分得分作为神经网络的输入信息, 建立了多神经元的神经网络感知器。 对农药残留检测的结果表明, 该方法可有效甄别农药残留和种类, 识别得到较好的分类效果。 总之, 该研究为蔬菜和瓜果表面的农药残留快速无损检测和识别提供了一条新途径。

A nondestructive testing based on visible/near-infrared reflectance spectroscopy was put forward for the common high pesticide residues of green plants in the wavelength range from 600 to 1 100 nm. Firstly, spectral features were extracted by wavelet transform from original spectral data. Secondly,the principal component analysis (PCA) was done in the further analysis of spectral characteristics. Thirdly, the two PCs were applied as inputs of artificial neural network, and a multi-neuron perceptron neural network was established. Finally, It was proved that the type of pesticide residues was effectively identified and showed by classification results. In short, the study provides a new approach to the detection of pesticide residues in vegetables and fruits.