A microscope auto-focusing system using the self-adaptive mountain-climbing search (SAMCS) method is introduced based on personal computer (PC) control. It mainly consists of four parts: the microscope, the digital camera to get the video images, the mechanical part of step motor and the computer to control the focusing process. The precision of the auto-focusing system is to some extent improved through high-resolution color images acquired by the digital camera as well as high subdivision of the step motor drive. An improved searching method--the SAMCS method is presented here. It can effectively improve the searching efficiency while guaranteeing the precision of the auto-focusing system. Based on the normal mountain-climbing search (MCS) algorithm, the SAMCS method takes full consideration of omnidistance concept and local extreme point influences. Thereby it can adaptively adjust the searching range according to different environmental conditions, and has quite good robustness. This feature mainly has two advantages. First, this method is much more exact than the normal mountain-climbing, which can not avoid local fluctuations. Second, it is much faster than the method of only using omnidistance searching to avoid local fluctuations. At the same time, we also take evaluation function and region selection into consideration to reach a faster and more accurate focusing result. And the experimental result demonstrates a good efficiency and accuracy.
[1]
Zheng Yu-zhen.
Optimized mountain-climb searching in auto-focusing
,
2005
.
[2]
Muralidhara Subbarao,et al.
Selecting the Optimal Focus Measure for Autofocusing and Depth-From-Focus
,
1998,
IEEE Trans. Pattern Anal. Mach. Intell..
[3]
K. Izumi,et al.
An Advanced Auto-Focus System for Video Camera Using Quasi Condition Reasoning
,
1990,
IEEE 1990 International Conference on Consumer Electronics.
[4]
Zhiliang Hong,et al.
Modified fast climbing search auto-focus algorithm with adaptive step size searching technique for digital camera
,
2003,
IEEE Trans. Consumer Electron..