Mechanical System Control by RGB-D Device

Abstract

Computer vision for control is a growing domain of research and it is widespread in industry and the autonomous vehicle field. A further step is the employment of low-cost cameras to perform these applications. To apply such an approach, the development of proper algorithms to interpret vision data is mandatory. Here, we firstly propose the development of an algorithm to measure the displacement of a mechanical system in contactless mode. Afterwards, we show two procedures that use a 3D camera as a feedback in control strategies. The first one aims to track a moving object. In the second one, the information gained from vision data acquisition allows the mechanical system control to ensure the equilibrium of a ball placed on a moving slide.