Research on Spectroscopic Determination and Classification of Diffuse Reflectance for Low-Altitude UAV Fuselage Materials

Abstract

The spectral reflectance characteristics of drone materials are one of the key factors enabling accurate spectral detection of drones. In this study, an experimental setup was independently constructed to measure the diffuse reflectance of drone fuselage materials. The reflectance spectra and their derivative spectral features of 15 different materials, including glass fiber, polypropylene, and polytetrafluoroethylene in various colors, were systematically analyzed. Based on this, machine learning algorithms such as Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbors (KNN) were applied to classify the aforementioned materials. The results show that the KNN algorithm demonstrated the best classification performance. Under the condition of a total sample size of 210 (training set: 147, test set: 63), the training set achieved an accuracy of 0.9731 through five-fold cross-validation, while the test set achieved a perfect accuracy of 1, indicating excellent model stability and classification precision. This research provides an important material study foundation for the spectral recognition and tracking of drone targets.