Three-dimensional fluorescence spectroscopy is a fast, nondestructive analysis method with good selectivity and high precision, which provides a foundation for the development of the current smart agriculture system. In modern agriculture, where agricultural information is fully perceived, it is still very difficult to quickly and destructively detect the internal chemical composition of soil, crops and agricultural products. Accurate determination of oil pollutants in water by using three-dimensional fluorescence spectroscopy technology can provide a basis for crop irrigation and is of great significance for improving agricultural benefits. The fluorescence spectrum analysis method is adopted to distinguish three kinds of mineral oil-gasoline, kerosene and diesel. In order to make the distinguishment more intuitive and convenient, a new identification method for mineral oil is proposed. The three-dimensional fluorescence spectra of the experimental dimension are reduced into two-dimensional fluorescence spectra. The concrete operations are as follows: adopting the method of end-to-end data matrix to constitute a large Ex image, and then figuring out the envelope curve, processing and analyzing the envelope image. Four factors, such as the ranges of excitation wavelength when the relative fluorescence intensity is greater than 0.5, the optimal excitation wavelengths, their kurtosis coefficients and skewness coefficients, are to be selected as the distinguishing feature parameters of mineral oil, and thus different kinds of mineral oil can be distinguished directly according to the feature parameters. The experimental results show that the proposed method has a high resolution for different kinds of mineral oil. Accurate and fast spectral data analysis methods can make up for the deficiencies of other agricultural information perception methods, provide a basis for the application of smart agriculture in many aspects and have a positive significance for promoting the comprehensive intelligent development of agriculture.
