Image-based visual servo for micromanipulation: a multiple-view and multiple-scale approach

In this paper, a novel approach to automatically align the micro part with both macro view camera and micro view microscope based on the eye-to-hand visual servo strategy is presented. First, we extend the existing visual servo techniques in micromanipulation to the case of multiple view vision sensors (one conventional CCD camera forms the perspective view, one microscope forms the microview) working in complimentary. Second, we show how to convert eye-in-hand strategy into an eye-to-hand based system, such that the velocity screw generated for both view are with respect to the same world frame. Third, the image processing and homography estimation procedures are presented to achieve a view-invariant presentation of three-dimensional features in the sense that it can be mapped to the corresponding image features without any exact knowledge about the camera parameters all the time during the positioning task.