Real-time precise 3D measurement of micro transparent objects using All-In-Focus imaging system

Recently micromanipulation has gained much interest for industrial and biological applications. To realize efficient manipulation of these applications, it is crucial to know the position of microobjects in 3D space. This means that both the micromanipulator and the target object should be detected and located in 3D space in a robust way. A number of different methods for 3D microobject positioning were proposed in the literature; however, either the object being manipulated or the micromanipulator is focused on with each having its own specific settings or assumptions. In our research, we obtained 3D position measurement of both the micromanipulator and target microobjects using an All-In-Focus imaging system. Because this imaging system is generally used for All-In-Focus visualization, it is necessary to overcome the noisy depth information provided by the system. This paper describes a robust 3D measuring technique to solve the noisy depth information problem and also how to locate the micro end-effector. In the experiment, 3D positions of a two-fingered microhand and glass microspheres were accurately obtained to automate the task of pick-and-place microobjects. Success rate over 70% was achieved. Real-time tracking of the microhand could also be achieved with success rate over 90%.