Designing and Implementing a Didactic Module of Artificial Vision for the Selection of Objects According to Colors and Morphological Characteristics

This present research describes the design and the implementation of a didactic module for the selection of objects, according to their colors and morphological characteristics using a robotic arm, through artificial vision, as well as the visualization of the process in a virtual environment. The project is equipped with a robotic arm that serves to select, capture, and locate nylon elements of different colors and shapes, placing them in different classification trays; it will also be visualized on a graphic interface through a monitor. Elements such as the Arduino data processing card, a 1920×1080 pixel HD camera to improve the visualization of the shapes and color of the element and a conveyor belt that will allow the movement of the elements have been used. Using the images obtained by the camera and processed in binary form in the Arduino, an automated control of the robotic arm that allows controlled movement and proper positioning in the corresponding sorting tray is obtained. The processing of the images uses a specific programming to make the edge detection and to obtain the points of the image that belongs to the border of the desired figure. Due to the imperfections of the image a process of filters in the image is followed as the scaling, the dilation of the image that allows suppressing the background of the figure and finally the erosion that helps to join or to obtain an outline to the desired shape. With the design and the construction of this didactic module of artificial vision, students and teaching staff will be provided with a better visualization of an industrial environment in automated inspection and quality control tasks, with the aim of improving the repetitiveness and precision obtained in a manual inspection process.