Optical simulation model of the diffuse reflectance near-infrared spectroscopy for predicting fresh maize quality

The optical properties of fresh maize tissues determine how light interacts with fresh maize cobs, which in turn affects the measured spectral signals and model accuracy. In this paper, a simulation model was developed to invert the optical properties of fresh maize cobs and evaluate the effects of different optical layouts on the accuracy of modeling predictions. First, the uniformity of detector irradiation at various distances (10 mm, 20 mm, 30 mm, 40 mm, 50 mm) and angles (30°, 45°, 60°) with different optical properties was analyzed using optical simulation methods. Then, the spectra of fresh maize cobs were collected at different light source angles and detection distances, and the spectral area polarization was calculated. Finally, the optical properties of the cob were estimated by establishing a link between irradiation uniformity and spectral area polarization, which resolved the distribution of light flux in edible maize cobs under different optical structures. The results show that the model of light transport mimicking the organizational structure of maize cob has been successfully simulated. The estimated optical properties of the cob are: absorption A=37%, transmission T=20%, and diffuse reflectance D=40%. This verifies that the accuracy and precision of the prediction model for the water content of fresh maize are best achieved under an optical structure with a detection distance of 40 mm and a light source angle of 45°. The establishment of the simulation model provides theoretical support for near-infrared detection of the intrinsic quality of fresh maize.