In　this　paper,　a　moving　objects　based　real-time　defogging　method　for　traffic　monitoring　videos　is　proposed.　Firstly,　dark　channel　prior　based　image　defogging　method　has　been　improved.　Then,　the　proposed　image　defogging　method　is　used　for　traffic　monitoring　video　defogging.　To　improve　the　processing　speed,　the　correlation　between　the　adjacent　frames　of　videos　is　exploited.　The　moving　objects　are　detected　using　adjacent　frame　difference　method.　The　frame　content　is　divided　into　moving　foreground　and　background.　Afterwards,　the　foreground　and　background　are　processed　with　different　defogging　manners　to　reduce　the　computational　complexity　of　defogging　processing.　Experimental　results　show　that　the　proposed　method　can　generate　a　good　defogging　effect　which　will　facilitate　the　subsequent　intelligent　traffic　analysis.　Furthermore,　the　proposed　method　is　fast　enough　to　process　the　standard-definition　videos　at　the　speed　of　26　frames　per　second　on　average.　©　2014　IEEE.