Spectral Imaging to Predict the Nutritional Quality of Rice
The quality of rice can be assessed by its phenolic and mineral content. A recent study shows how to predict the nutritional quality of rice using mathematical modeling and spectral imaging. Innovative prediction methods have been proven to aid organizations in breeding crops with higher quality and nutritional levels.
Why does the quality of rice matter?
Over 50 percent of the population depends on rice, and most is consumed in developing countries (fao.org, 2023). The most commercialized and accessible type of rice is the white variety; however, this typology is perceived as a mono-crop due to a lack of nutrients. Consumers treat rice as a food staple in their diet, even though it is full of carbo-hydrates – its overconsumption might lead to malnutrition.
Rice is a crucial part of most of the population’s diet; hence food organizations collaborate to popularize the breeding of more nutritious and healthy types of rice.
Pigmentation of rice as a determinator of its quality
Pigmented rice is characterized by red and purple colors. The natural coloring appears from the presence of phenolic compounds like phenolic acids, flavonoids, and anthocyanins. Those compounds have shown richness in nutrients resulting in antioxidant, anticancer, and other beneficial properties. Furthermore, pigmented rice appears to have minerals such as iron, magnesium, potassium, and more.
Naturally colored rice is gaining popularity among consumers due to its undeniable positive effects on human health. Research on further correlation between rice pigmentation and its nutritional quality is being increasingly undertaken, in order to understand the benefits pigmented rice brings to human nutrition.
A spectral imaging study on the prediction of nutritional quality of rice
As part of the CGIAR’s Research Initiative on Market Intelligence, a study was conducted with the use of multispectral imaging to predict the quality of rice in relation to the content of bioactive and minerals. The International Rice Research Institute (IRRI) supplied the researchers with various rice samples in terms of their type and harvest time. The focus of the study was to develop a prediction model for total phenolic content (TPC), total flavonoid content (TFC), total anthocyanin content (TAC), as well as mineral content in each sample.
The researchers chose to use the VideometerLab for the spectral phenotyping of the rice kernels. Random forest (RF) and artificial neural network (ANN) were used as models for predicting the nutritional quality of rice. The results of the study show that it is possible to predict the nutritional quality of rice by using spectral imaging technology and mathematical modeling. The accuracy level of ANN and RF models is 87.6% and 75.43%, respectively. The researchers illustrated that the prediction models can be further used for rice breeding to determine phenolic and mineral content based on the sample color and biochemical indicators.