Research Of Autofocus Technology Based On Image Processing

The paper introduces the development and application of auto-focusingtechnology, analyzes the principle of autofocus, and then studies auto-focusingmethods based on image processing on the basis of summarizing variousauto-focusing algorithms. It takes one kind of low-cost and special video processorTMS320DM6437with DaVinCi technology as developing platform to performautofocus based on image processing.Image clarity evaluation function is the key point of autofocus by using imageprocessing methods. Through analyzing various kinds of evaluation functions, itproposes an evaluation function based on gradient threshold to improve the real-timeand anti-noise performance. In addition, it analyzes focusing window selected andfocusing searching strategy which have an effect on autofocus, and points out theimportance of selecting the focusing window adaptively. Aiming at the case of thefocusing curve may exist local extremums, it proposes an improvedmountain-climbing algorithm to search the focusing extremum to decrease the effectof local extremums.At last, it accomplishes auto-focusing program on the CCS3.3softwareplatform and builds hardware platform to verify autofocus based on imageprocessing on the TMS320DM6437evaluation board.