一种改进的自动调焦爬山搜索算法

26320105JournalofKunmingMetallurgyCollegeVol26No3May.2010:2010-04-20:(1983-),,,,doi:10.3969/j.issn.1009-0479.2010.03.008牟宏鑫1,吴庆畅1,张翠1,吴诚2(1.,650500;2.,650223):图像法自动调焦的一个关键问题在于图像清晰度评价函数的选取,自动调焦的过程就是求取图像清晰度评价函数最大值的过程针对普通的爬山搜索算法易受到调焦评价函数局部极值干扰而不能准确调焦的问题,提出了一种改进的爬山搜索算法该算法以连续3次的调焦函数值的变化情况作为确定调焦方向的判据,有效地避免了局部极值的干扰:自动调焦;图像清晰度;调焦评价函数;爬山搜索算法:TN91173:A:1009-0479-(2010)03-0032-04AnImprovedMountainclimbSearchingAlgorithminAutofocusingMUHongxin1,WUQingchang1,ZHANGCui1,WUCheng2(1.FacultyofElectricalandInformationEngineering,YunnanUniversityofNationalities,Kunming650500,China;2.KunmingPhysicsResearchInstitute,Kunming650223,China)Abstract:Akeyissueoftheimageautofocusingmethodistheselectionofthefocusingevaluationfunction.Theprocedureofautofocusingistogetthemaximumofthefocusingevaluationfunction.Ordinarymountainclimbsearchingalgorithmisapttobedisturbedbylocalextremum.Tosolvethisproblem,animprovedmountainclimbsearchingalgorithmisproposed.Thisarithmeticiseffectivelyavoidthedisturbancebyusingthreeconsecutivechangesofthefocusingevaluationfunctionasthecriteriontoconfirmthefocusingdirection.Keywords:autofocusing;imageresolution;focusingevaluationfunction;mountainclimbsearchingalgorithm019,1898,[1],,,,,,,,,[2],,[4],;,,,,Brenner,[3]:F(K)=MN[I(x+2,y)-I(x,y)]2(1),I(x,y)xyF(K),11[5],,,1,AB,BA,AB,,,,,,,,,1CD,P,221,,,,,,2,P122,59,RobertsTenenGradLaplaceBrenner3333,,,:3,,,,,2,34,3,;,(4A),(4B),235:forj1=1:m(i1)forj2=1:n(i1)-2sum2=sum2+(double(I{i1}(j1,j2+2))-double(I{i1}(j1,j2)))^2;endendhe2(i2)=sum2;sum2=0;v=length(he2);ifv4forj=1:v-434201056if(he2(j+1)he2(j)&he2(j+2)he2(j+1)&he2(j+3)he2(j+2)&he2(j+4)he2(j+2))jd1=j+1;endendendifjd10breakendend24MATLAB,3,Brenner,,,6773,,,,,,,,,,,:[1],,.[J].,2007,37(11):1213-1215.[2],,.[J].,2003,30(2):53-55.[3],,.[J].,2003,30(2):53-55.[4],,,.[J].,1999,20(4):85-87.[5],,.[J].,2007,21(1):13-16.[:]353,,,: