Generally, clarity of images or focal value (FV) varies
in hill-shaped manner with respect to lens motion in a focus
region. However, the FV variation in the far-off focused region
is usually subjected to disturbances and the image signals may
not be informative for focusing control. Therefore, challenges
exist if focusing control under microscopic views for a large
depth of lens displacement is needed. The challenges are
primarily due to lack of a method to differentiate image FVs at
far-off focused regions.
This paper proposes a method to integrate regional
monotones of FV variation. A focusing control algorithm with
step-size adaptation is derived based on the above monotonous
properties to gain speed and reliability. Information about the
regional monotones is acquired and used for checking and
correcting inconsistency in the motion direction of lens caused
by image disturbances. This inconsistency check is significant on
enhancing the robustness of focusing control.
Four case studies are illustrated to demonstrate the feasibility
of the proposed algorithm and the effects of design parameters.
A response time of 0.6~1.2 sec with perfect reliability is achieved.
This speed is relatively sufficient for many industrial
applications, such as LCD panel laser repairs.