Research And Implementation Of Auto-focusing Algorithm For Integration Camera

In recent years, integration camera became the core equipment in TV monitoring system.Integration has some Characteristics, such as simple structure, convenient, small, simple controlsystem, so it was used all over the place in monitoring field. And the development of integratedcamera mainly thanks to the development of auto-focusing technology, the technology ofauto-focusing is a key technology in the process of get image from various imaging system. Withthe development of modern computer application and digital image processing technology,auto-focusing has entered a new stage, more and more auto-focusing algorithm is based on thetheory of image processing. So the research of auto-focusing algorithm has important practicalsignificance.This paper first introduced the auto-focusing of basic principle, characteristic and classification,and then this paper had a detailed study on some core technology of auto-focusing: focusmeasurement function, focus regional selection algorithm and peak searching algorithm. Finally, westudied on the design of hardware and software for auto-focusing system. On this basis, using theFPGA development platform to realize auto-focusing function. The process of auto-focusing is theregion of interest to focus, focus measurement function is judge defocus level for region of interest,peak searching algorithm is compared the values of focus measurement function to move around thecamera, and lens moved to the position of the best image quality, finally completing the process ofauto-focusing.Focus measurement function of auto-focusing was analysed in this paper, focus measurementfunction was introduced from time domain analysis to method of frequency domain, and thenfocusing on the FSWM filter algorithm (weighted median filter algorithm). Finally, we test focusmeasurement function of time domain analysis and focus measurement function of time domainanalysis was test and analysed its performance. The experimental results show that FSWM filteralgorithm has the very good anti-noise ability, finally this algorithm is applied to the design of theauto-focusing system.This paper introduced the importance and the main principles of focus regional selectionalgorithm for auto-focusing. All sorts of one dimension, two dimensional focus regional selectionalgorithm are described in this paper. The method based on golden section was introduced to selectthe focus window in our auto-focusing system, peak searching algorithm was used for searchingand position focus. This paper discussed the Fibonacci search algorithm, the curve fitting searchalgorithm, mountain climbing search algorithm. And the improved mountain climbing searchalgorithm was put in this paper, this algorithm is applied to the design of the auto-focusing system.At last, this paper introduces the hardware and software research and implementation process of the auto-focusing system, hardware mainly used S2C’s Virtex-5 TAI Logic Module prototypetest platform as the core, and put the system into the optical lens part, image acquisition module,image processing and master control module, motor driver modules. Software mainly introducedthe operation process of the auto-focusing system. The design of the auto-focusing system was testand analysed by us, the results show that this system can accurately, stable completeauto-focusing.