Performing tasks in a coordinated manner by means of a team of robots is a topic of great interest and allows to improve the results compared to the single-robot case. The current research project focuses on this particular field and proposes the need to use different vision systems distributed along the mobile agent network that gather a precise and complete description of the environment. To cope with the proposed goals it will be necessary to tackle with different research lines, in consequence, we are working on the following subjects: Cooperative map building and localization using particle filters, Visual landmark modelling: Improving data association in visual SLAM, development of cooperative exploration strategies using the information provided by each robot, cooperative reconstruction of environments using appearance based methods.