Monitoring fungal growth on brown rice grains using rapid and non-destructive hyperspectral imaging.

This research aimed to develop a rapid, non-destructive, and accurate method based on hyperspectral imaging (HSI) for monitoring spoilage fungal growth on stored brown rice. Brown rice was inoculated with a non-pathogenic strain of Aspergillus oryzae and stored at 30 °C and 85% RH. Growth of A. oryzae on rice was monitored using viable colony counts, expressed as colony forming units per gram (CFU/g). The fungal development was observed using scanning electron microscopy. The HSI system was used to acquire reflectance images of the samples covering the visible and near-infrared (NIR) wavelength range of 400-1000 nm. Unsupervised self-organizing map (SOM) was used to visualize data classification of different levels of fungal infection. Partial least squares (PLS) regression was used to predict fungal growth on rice grains from the HSI reflectance spectra. The HSI spectral signals decreased with increasing colony counts, while conserving similar spectral pattern during the fungal growth. When integrated with SOM, the proposed HSI method could be used to classify rice samples with different levels of fungal infection without sample manipulation. Moreover, HSI was able to rapidly identify infected rice although the samples showed no symptoms of fungal infection. Based on PLS regression, the coefficient of determination was 0.97 and root mean square error of prediction was 0.39 log (CFU/g), demonstrating that the HSI technique was effective for prediction of fungal infection in rice grains. The ability of HSI to detect fungal infection at early stage would help to prevent contaminated rice grains from entering the food chain. This research provides scientific information on the rapid, non-destructive, and effective fungal detection system for rice grains.