Identification of the Pesticide Fluorescence Spectroscopy based on the PCA and KNN

The organic pesticide can emit fluorescent light when it is excitation by the ultraviolet ray, carbonates pesticide can identification by the three-dimensional fluorescence spectroscopy technology. Statistics based on the apparent feature is limitation to identification of the complex Pesticide Fluorescence Spectroscopy. In order to realize identification of the diversity pesticide which Fluorescence Spectroscopy have more overlapping, pesticide spectroscopy is compression and dimensionality reduces by Principal component analysis (PCA). Spectroscopy character of the various objects is extracted. KNN classify method is combined to realize the sort identifications of carbaryl, mipcin, furadantin and aldicard are implementation. The sorting result is visualization by parallel coordinate chart. Experiment result indicates this method is base on utility information and effectiveness dimensionality reduction dispose to high dimensional spectroscopy information. Sorting speed is much higher. The identification rate reaches 97 percent. The result of identification is better.