Innovación tecnológica para la reducción de la heterogeneidad del aguacate ''Hass'' cosechado para exportación

Abstract

For ''Hass'' avocado exportation is required high quality standards, such as fruit maturity. However, the harvested fruit presents heterogeneity of ripeness, requiring a solution that allows identifying the optimal time of harvest. The optimal time is associated with the physiological maturity. And the dry matter (D.M.) content is used to measure the maturity. But, the nondestructive methods to predict its content are very expensive or are not operative. So, we present a solution to identify if a fruit has physiological maturity on the tree. The solution involves a method and a device with two use cases: classification and system parameterization. The classification use case use a support vector machine trained to classify a vector of texture descriptors calculated from an image of the fruit on the tree. The image is acquired in RGB format, without compression and with homogeneous spatial resolution. he image is pre-treated with a contrast limited adaptive histogram equalization and a conversion to an HSV color space. A segment of the HSV image is used to calculate an optimized vector of texture descriptors. The parameterization use case is to optimize the vector of texture descriptors. This vector is a set of textures descriptors with high efficient cross-validation classification and lower computational cost. The device implements the method and must perform the classification in real time, with portability and ruggedness. The main contributions of the process of research, development and innovation are: This research shows that the fruit has a heterogeneous maturity. For that we test all the fruits of a tree in a harvest season, presenting a variation among fruits upper to 20% of D.M. . we prove the viability of the method in a relevant environment, which a classification efficiency of 98.2% that supports the patent application of the proposed solution. And if a farmer implements the invention, he will have an increase of income in 91.4%.