Ergonomy of vision through television

We consider manipulation tasks of objects in a 3D scene using a telemanipulator, in which the operator views the scene on a TV screen. We study the influence of lighting strategies on the performance of the manipulation tasks, both in terms of object recognition and localization, and in the time it takes to execute grasping and placement operations. The results show that the use of multiple lights in different geometrical setups is at best as effective as the use of structured light. The projection of 2 grid patterns from perpendicular directions led to the best overall performance.