Plant Phenomics / 2021 / Article / Fig 3

Research Article

Classification of Rice Yield Using UAV-Based Hyperspectral Imagery and Lodging Feature

Figure 3

Processing workflow of hyperspectral and RGB images and dataset construction. Raw images from heading stage were adjusted for several types of distortions to obtain corrected data. Reference spectra were extracted, and noise and background were removed. After grid cell and subsequent ROI selection, random pixels were used to calculate vegetation indices. Lodging statistics from mature stage were extracted and combined with the spectral data to provide training and testing set.