Autofocus performance realization using automatic control approach

Modern autofocus systems are known for their fast response speed and optimized routines of the image processing. Nevertheless, the bottleneck of the autofocus system development is the transformation of the technical specification into the design rules that strictly take in account the specifications and introduce a guarantee of the best performance. In this paper we propose a new approach in autofocus systems design that uses some linear control theory results that allow to obtain the optimal feedback gain in the sense of quadratic sharpness hypothesis. The method proposed in this work allows to improve the existent autofocus systems that are based on the sharpness autofocus using the optimal feedback gain. It is shown that the method is sufficiently generic and can be integrated into an image signal processing unit of a modern camera.