A team of researchers from the University of Granada (UGR) has developed a method to determine the ripeness of peppers using hyperspectral imaging and machine learning.
The research was conducted with bell peppers, which are known for their thick walls. The new technique makes it possible to determine the firmness of the peppers, a property related to their ripeness and, therefore, to their acceptance on the market.
The new method was developed by a team of researchers from the Color Imaging Lab of the UGR, with the collaboration of the Department of Analytical Chemistry and the Mabe Fruit and Vegetable Cooperative of Almeria, to prevent overly ripe peppers from being packaged.
They achieved this by using hyperspectral images, a non-invasive technique used for the detection and visualization of quality characteristics in fruits and vegetables.
The scientists analyzed the spectral reflectance of peppers in the visible and near-infrared range to determine the ripeness of peppers from three different crops by identifying the spectral bands that best represent the degree of ripening of peppers.
Researchers developed a realistic scenario similar to a conveyor belt system where peppers are evaluated with four sorting algorithms to predict their ripeness.
The success rate of the algorithm used for this classification has been above 90%. This system improves how the shelf life of peppers is evaluated and can help guarantee the quality of the product is better for customers.
According to the UGR, the system provides an effective and practical solution for determining the ripeness of peppers and allows vegetable growers and packing houses to improve harvest management and reduce product waste. In addition, the flexibility of the system allows companies to adjust the number of spectral bands according to their budget and the type of product they wish to analyze.
Source: ugr.es