Target　detection　of　oil　pipeline　for　the　perspective　of　Unmanned　Aerial　Vehicle　(UAV)　is　one　of　the　important　topics　of　UAV　oil　pipeline　patrol.　The　existing　methods　have　limitations　for　the　image　with　complex　background,　and　can　only　identify　the　location　of　the　pipeline,　but　cannot　obtain　the　specific　geometric　information　and　width　characteristics　of　the　oil　pipeline.　In　this　paper,　an　oil　pipeline　detection　and　extraction　algorithm　based　on　UAV　pipeline　patrol　is　proposed.　Wavelet　multi-scale　edge　detection　is　used　to　improve　the　robustness　of　pipeline　edge　detection　in　complex　background.　Based　on　the　stroke　width　transform,　the　specific　geometric　information　of　the　oil　pipeline　is　obtained.　The　pipeline　position　information　can　be　obtained　based　on　Hough　linear　detection.　On　this　basis,　the　pipeline　detection　algorithm　process　is　proposed　and　tested.　The　method　can　obtain　more　information　about　the　pipeline　more　accurately,　which　is　the　necessary　prerequisite　for　locating　pipeline　deformation　in　the　process　of　UAV　oil　pipeline　patrol.　©　2021　ACM.