After　comprehensively　studying　OSTU　method　with　maximum　variance　between　two　classes　and　the　genetic　algorithm　(GA)　principle,　an　OSTU　method　based　on　the　self　adaptive　genetic　algorithm　(SAGA)　is　proposed　for　image　segmentation.　In　this　paper,　the　OSTU　method　is　optimized　in　the　GA　based　on　the　encoding,　selection,　crossover　and　mutation　operations.　The　experimental　results　show　that　by　using　SAGA　method,　the　efficiency　of　getting　the　optimal　solutions　can　be　improved,　the　premature　problem　can　be　avoided　to　a　great　extent　and　finally　the　satisfying　solution　can　be　achieved.　The　image　segmentation　results　show　that　the　OSTU　method　based　on　the　SAGA　is　not　only　quicker　in　computing　speed,　but　also　higher　quality　in　image　segmentation.　On　the　basis　of　the　better　image　segmentation　result,　the　target　position　in　the　CCD　image　can　be　extracted　more　accurately　in　the　centroid　method.　©　2007　IEEE.