| With the rapid development of the electronic information and the network technology,the digital image and video technologies are highly demonded by many fields and industries.Auto focus is one of the key technologies in computer vision,image-forming systems and various of precision instruments,the image and video capture devices with a auto focus feature play a important role in consumer electronics,security monitoring,national defence and industrial production.Most of the precision measuring devices are required to equip with auto focus technology,which affect the measure result greatly.With the utilization of digital image-forming devices,the recent years'researches concentrate mainly on the methods based on digital image processing,with abundant accomplishment.However,due to the complexity of image-forming conditions and targets,there are lots of defects existing in domestic auto focus technologies,such as low speed and low success rate,etc.Meanwhile,the information revolution has produced the digital signal processing technology,and impelled its development.With its programmable feature,DSP(digital signal precessor) has become a increasingly important processor.In recent years,DSP has been perfecting in intelligent video monitoring,and taking the place of traditional unintelligent systems.The image evaluation algorithm and motor searching strategy are two core parts in auto focus technologies based on image-analysing.Different with other theoretical studies,this paper has built a hardware experiment platform for the research.Firstly,the paper gives a deep analysis of various image evaluation algorithms, with the conclusion that different algorithms work with different performances,and we choose a best evaluation algorithm and put into use.Secondly, the paper offers an improved vision of hill climbing search algorithm,which shows a better performance.At last,from the experiments,we has discoveried the factors that affect focus result. |
