Mobile robot vision navigation strategy based on Pixy2 camera

摘要

This paper aims to present an efficient navigation method for collecting tennis balls using a visual sensor, specifically the Pixy2 camera, which, in addition to the sensor, also has a high-quality processor built in. In order to achieve this goal, a navigation algorithm based on three situations encountered in practice when detecting a tennis ball was developed and the information received from the Pixy2 camera, that is width, height and two coordinates on the x-axis and y-axis was processed. Basically, two hypotheses were considered, namely: area decides the order and object coordination defines the choice of navigation strategy. Using a square consisting of 100x100 [mm] gray squares, the navigation method was validated based on a visual sensor, by studying five individual and representative cases of tennis ball location. For each case, the four data provided by the sensor were collected and then processed in order to highlight the two hypotheses initially established.