On Auto Focus Theory And Technology Of Digital Image Systems

Digital imaging system (DIS) is widely used in many practical areas such as industrial process control, field monitoring and military affairs, auto focus is one of the most important technologies in DIS. The traditional auto focus technologies depend mainly on some special assistant devices which are used to measure distance, find focus. As a result, these methods have some disadvantages such as high cost and low sensitivity, particularly do not satisfy the demand of real-time system.This thesis is originated from an actual project to build a Camera Assistant FocusSystem (CAFS), and mainly includes the research and development work as follows:1. After sufficiently investigating on image preprocess, focus-windows selection strategy, feedback control algorithm and so on, the author proposes the basic principle of CAFS, further presents a new algorithm for image definition evaluation, which is the crucial part of auto focus technology.2. The new algorithm is compared with the existing methods of auto focus through theoretical analyses and simulative experiment. And it is proved to be more effective due to its speedy calculation, and better real-time property. Meanwhile the new algorithm has been applied in CAFS made by us.3. USB2.0 Image Capture Card as the key devices of CAFS is developed, which includes its firmware designed with Keil C tools, as well as its Windows device driver designed with DDK and Driverworks tools.Benefited from satisfaction of industrial real requirement, the CAFS holds ideal stability and practicability, specially improves the efficiency and product quality when focus-fixed cameras are manufactured massively. Finally, the author prospects the development trend and main research topics of auto focus technology in the future.