Development of an artificial vision system for the detection of dents in tuna cans, Trujillo

Abstract

In this work, the implementation of an artificial vision system is presented for the detection of dents and imperfections in unsealed cans without content that are presented or may be presented in canned fish, when handled and transported they suffer knocks that compromise its structure and affect its presentation. The main objective is the development of an algorithm that allows efficiently detecting flaws in the edges and the general structure of an aluminum tuna can without product and without sealing and the specific objectives are: to determine the appropriate techniques for the detection of edges defective by applying a simple programming language and establishing the characteristics and sequence of the logical process to follow at the time of starting the detection, 5 trials were carried out with 10 tests each, where 10 cans of tuna were entered and it was simulated using a Python code editor, with which the operation of the machine and tests to measure efficiency would be carried out. Finally, it was concluded that the efficiency percentage was 90.4%, this being very beneficial for a company in its production and quality area.