抄録
This paper tackles challenges of general realtime position based visual servoing without using model information. These are mainly speed limitations of suitably affordable vision processing system, generally reliable features and faithfully tracking the target to keep it in field of view. Feature detection is not only time consuming but also depends on actual feature parameters. The proposed autotuned feature detection enables the system in reaching a higher speed by reducing processing time and also has the advantage of easiest portability to another vision processing system. Simulations with real images and experiments have been conducted to prove the effectiveness of the method.
