Non-destructive modeling of mango geometrical attributes assessed using image processing technique for mass and volume measurement

Image based modeling of mangoe geometry

Abstract

The mechanization of post-harvest operations and the management of mangoes are crucial for minimizing post-harvest losses. Economic factors, the quality of the fruit, and the reputation of mango growers and exporting countries are significantly impacted by traditional sorting and grading techniques. Therefore, obtaining quick and accurate geometrical data is essential for designing equipment for sorting, grading, loading, unloading, and packaging mangoes after harvest. This study focuses on capturing images using a near-infrared (NIR) camera, processing those images, and analyzing external dimensions (length, breadth, and thickness) using LabVIEW software. The geometrical attributes obtained were then utilized for non-destructive modeling of mango mass and volume. The correlation strength, indicated by the R value of the calibration regression equation, ranged from 0.634 to 0.932 for mass, and from 0.624 to 0.805 for fruit volume. The findings indicated that the best mass models were based on the length, arithmetic mean diameter, and all three external dimensions (length, breadth, and thickness). Conversely, the volume models that provided the best fit were based on the length and breadth of the fruits.