Diseño e implementación de un sistema de detección de irregularidades presentes en una pieza de cuero post-curtido basado en visión artificial para la Empresa Curtiduría Tungurahua S.A.

Abstract

In the present research work, an artificial vision-based detection system was designed and implemented to identify irregularities in a piece of leather at tannery CURTIDURÍA TUNGURAHUA Co. To carry out the vision system, the heuristic and experimental methods were used, the system was based on a support structure for the Nikon D3200 camera and 2 LEDEX flood lamps with a platform where the piece of leather lies. The modeling of structures was made in SolidWorks thus obtaining the measure for the installation of the different components. The system obtains the image by means of using the DigiCamControl program, the image processing was developed in the Python Opencv programming language to obtain an image of the leather and so detecting existing damage and validating optimal performance and then determining efficiency when detecting irregularities. It is concluded that the system can detect existing irregularities in the leather with an effectiveness between 97% to 99% with a 40% of time saving, acceptable by experts in the classification area. It is recommended for this application to get a high-end robust camera for its higher pixel count with a focus angle greater than 110 °.