Infrared Video Background Division Technology Research

Along with the development of infrared image processing technology, infrared technology inguided, automatic control and so on many domains a wide range of applications. Although throughinfrared imaging system for the video image with the noise of generation, get video images of theeffect is not good, And sometimes we want vision looks very fuzzy, and carry on the processing isalso up trouble, make our research difficulty also add greatly, but, because infrared imagingequipment can through dust, fog to detect obstacles; And imaging system has higher concealment,long-distance detection, etc. So using infrared technology acquisition image, and the processinghas important effect, however, for the video image segmentation for operation, is the follow-up ofthe identification and tracking the necessary link, then the infrared video image segmentationalgorithm research has important significance and value.This paper first introduced the infrared imaging process causes of the noise and theclassification, in order to study the process of getting clear video image, we can think of noise isuseless information. Through the gray level correction, histogram transformation, the imagesmooth the image preprocessing operation get clear images.Secondly introduces some of the video segmentation algorithm, and points out their shortage,frame difference algorithm is not suitable for extract complete goal, single gaussian backgroundmodeling is not suitable for complicated background environment.Finally, this paper brings forward based on the time domain target extraction algorithm,Through the time domain video sequence image, the division of the effective, then, for the variance,mean, we want to get the background image, using maximum likelihood operation, open operationand so on obtains the segmentation result, proved by the experiment, the algorithm can adapt to thecomplex environment and realizing the extraction of target motion.