Signal-to-Noise Ratio Contributors and Effects in Proximal Near-Infrared Spectral Reflectance Measurement on Plant Leaves

摘要

An extensive study was performed recently on Jacobaea Vulgaris to characterize the plant leaves using the hyperspectral reflectance. Unfortunately, a poor signal-to-noise ratio (SNR) in the spectral data was observed thus making the results less valuable. This paper examines the consequence of using uncontrolled lighting, i.e., solar illumination in the field applications as well as the effect of scattered light on shaded regions of the image. It is shown that a significant reduction of SNR can be achieved by taking these factors into account prior to the processing of the hyperspectral data. Specifically, the standard uncertainty of the spectra for Jacobaea Vulgaris hyperspectral images can be reduced from an average of 50% to better than 20%.