Mass estimation of mangoes by processing of white background grayscale images.
Mata M. E. R. M. C., Duarte M. E. M., Borges F. F., Barbacena I. L.
Author Affiliation: Agricultural Engineering Department, Federal University of Campina Grande, Campina Grande, PB, Brazil.
: 937-942
Abstract : The postharvest grading of mangoes normally includes mass measurement using expensive dynamic weighing systems to group similar fruits in the same package. Small producers cannot buy systems like these which are normally complex to install and operate/maintain. This work describes a study for grading of mangoes (Mangifera indica L., 'Tommy Atkins') based on image processing. It aims to apply image processing, afterwards, on a low cost postharvest fruit selection and grading system. A hundred mangoes were acquired at the local market of Joćo Pessoa (PB), northeastern of Brazil, and divided into two 50 units groups, randomly. A SonyTM P7 digital camera was fixed at the center top of a closed polystyrene box (50×50×50 cm), internal walls covered with white paper. Two PL(20W) lamps illuminated the mango at the center bottom of the box. Top view picture of the fruit, at rest, was taken. An offline MatlabTM software calculated the top area of the fruits using a new threshold value y=x+x.(m/k), where x: Otsu's threshold; m: mean intensity of the image previously segmented using Otsu's algorithm; k: empirical constant, ambient light dependent. All the fruits were weighed using a digital scale and the first group was used in order to find (projected top area × actual mass) best linear relation: m(g)=-108.7+0.002649.A(pixels2). The other 50 fruits images were used aiming to validate the equation found. Result showed a highly linear correlation between the top area and the measured masses (R2=0.96, SD=15.7 g) for the first group. It was possible to estimate, by using the area of the top images, the masses of the mangoes belonging to the second group. The (predicted × actual) mass values showed R2=0.94, SD=19.8 g. Research is being carried out in order to improve the grading system on a conveyor belt, indicating also the possible use with other horticultural products.