Diseño e implementación de un prototipo clasificador de huevo de gallina basado en las imperfecciones de la cáscara aplicando visión artificial.

Abstract

A hen egg classifier prototype was designed and implemented based on the imperfections of the shell applying artificial vision. The prototype focuses its operation on a Minicomputer raspberry Pi3, which connected to a webcam Genius F100 acquires images of the various faces of the egg applying the technique of Candling along whit an array of mirrors, ensuring the full coverage of Area. The acquired images are processed within the raspberry by implementing an algorithm written in the Python programming language using the Vision libraries offered by OpenCV. Statistical methods and artificial vision are used to detect the anomalies present on the surface of the egg shell. Culminate with the processing the eggs are transported through a series of mechanisms, where they will be classified in different lanes with the help pf actuators servomotors, the egg that present anomalies will be the first to be separated giving way to that the eggs that do not present them making use of extensiometric gauges respectively conditioned, will be weighed and classified according to their size, by means of an algorithm of comparison of weight programmed in the Arduino card complying with the established in INEN standards. The control and classification logs are stored in a database, which can be viewed using a man-machine interface (HMI). From the tests carried out with the prototype, it was obtained as a result a time saving of approximately 60% with respect to the manual classification mode. It is concluded that the implementation of the prototype improves the quality of the product and allows a greater efficiency of the process carrying a daily production control in a database stored by the system. It is recommended to follow the instructions of the authors to guarantee the correct functioning of the prototype.