Hartmann　wavefront　sensor　is　applied　widely　in　adaptive　optics　systems.　Considering　the　real-time　performance　requirements　and　processing　amount　of　Hartmann　sensors　in　computing　the　centroids　of　points,　this　article　proposes　an　FPGA-based　real-time　image　processing　and　centroid　extraction　system.　The　system　adopts　FPGA　as　the　core　processor　and　uses　the　connected　domain　labeling　algorithm　and　the　gray　centroid　method　to　complete　the　centroid　calculation.　In　addition,　the　calculation　results　are　output　simultaneously　through　multiple　DAs　to　drive　the　device　under　test　to　achieve　image　wavefront　restoration.　At　the　same　time,　the　image　data　is　uploaded　to　the　host　computer　in　real-time　through　the　Ethernet　interface.　The　system　has　excellent　performance　in　real-time　image　processing,　image　output　quality,　and　system　control　stability.　Experiments　show　that　it　is　an　efficient　technical　means　of　providing　real-time　wavefront　processing　measurement　for　the　Hartmann　sensor　in　various　application　scenarios,　such　as　adaptive　optical　correction.　©　2023　IEEE.