3D computer-vision system with variable spatial resolution

A computer-implemented method for controlling movements of a robotic arm comprising an end-effector, the method comprising:

• determining, based on visual information provided by a low-resolution three-dimensional (3D) camera module, an initial pose of the end-effector, wherein the low-resolution 3D camera module is part of a dual-resolution computer-vision system comprising the low-resolution 3D camera module and a high-resolution 3D camera module, wherein the low-resolution 3D camera module and the high-resolution 3D camera module are arranged in such a way that a viewing region of the high-resolution 3D camera module is located inside a viewing region of the low-resolution 3D camera module;
• determining an intermediate pose of the end-effector based on the initial pose and the viewing region of the high-resolution 3D camera module;
• controlling, by a robotic controller, movements of the robotic arm to move the end-effector to the intermediate pose such that a component held by the end-effector enters the viewing region of the high-resolution 3D camera module;
• determining a pose of the component based on visual information provided by the high-resolution 3D camera module;
• determining a final pose of the end-effector based on the pose of the component;
• and
• controlling movements of the robotic arm to move the end-effector to the final pose.