Underwater　Surveillance　Video　(USV)　plays　important　role　in　the　research　and　development　of　marine　resources.　The　harsh　underwater　environment　means　that　USV　can　only　be　transmitted　over　hydroacoustic　channel　with　narrow　bandwidth.　Region　of　Interest　(ROI)-based　rate　control　algorithm,　which　automatically　adjusts　the　code　rate　to　accommodate　bandwidth　changes　based　on　user＇s　interest.　So　the　ROI　Extraction　Algorithm　(ROI-EA)　is　a　key　process.　But　the　existing　ROI-EA　is　mainly　aimed　at　conventional　video.　It　does　not　take　into　account　the　characteristics　of　USV.　Thus,　there　is　still　room　for　improvement　in　its　performance.　Therefore,　an　ROI-EA　for　USV　is　proposed　in　this　paper.　The　characteristics　of　USV　was　explored　first.　Then　a　underwater　image　enhancement　algorithm　based　on　improved　gamma-corrected　was　proposed.　Finally,　the　proposed　ROI-EA,　according　to　the　characteristics　of　USV,　was　demonstrated.　We　also　compared　the　performance　of　our　method　with　existing　ROI-EAs.　The　results　show　the　superiority　of　our　algorithm.　©　2023　IEEE.