谷速软件网matlab有偿编程matlab伯恩森本地图像阈值源码程序Word格式文档下载.docx
《谷速软件网matlab有偿编程matlab伯恩森本地图像阈值源码程序Word格式文档下载.docx》由会员分享，可在线阅读，更多相关《谷速软件网matlab有偿编程matlab伯恩森本地图像阈值源码程序Word格式文档下载.docx（18页珍藏版）》请在冰豆网上搜索。
%----------
%TherearedifferentwayshowtoperformmeanfilteringinMATLAB.
%AneffectivewayforsmallneighborhoodsistouseIMFILTER:
%I=imread（'
eight.tif'
）;
%meanFilter=fspecial（'
average'
[33]）;
%J=imfilter（I,meanFilter）;
%figure,imshow（I）,figure,imshow（J）
%However,IMFILTERslowsdownwiththeincreasingsizeofthe
%neighborhoodwhileAVERAGEFILTERprocessingtimeremainsconstant.
%Andonceoneoftheneighborhooddimensionsisover21pixels,
%AVERAGEFILTERisfaster.Anyway,bothIMFILTERandAVERAGEFILTERgive
%thesameresults.
%Remarks
%-------
%TheoutputmatrixtypeisthesameasoftheinputmatrixA.
%Ifeitherdimesionoftheneighborhoodiseven,thedimensionis
%roundeddowntotheclosestoddvalue.
%Example
%J=averagefilter（I,[33]）;
%SeealsoIMFILTER,FSPECIAL,PADARRAY.
%ContributedbyJanMotl（jan@motl.us）
%$Revision:
1.2$$Date:
2013/02/1316:
58:
01$
%Parameterchecking.
numvarargs=length（varargin）;
ifnumvarargs>
2
error（'
myfuns:
somefun2Alt:
TooManyInputs'
...
'
requiresatmost2optionalinputs'
end
optargs={[33]0};
%setdefaultsforoptionalinputs
optargs（1:
numvarargs）=varargin;
[window,padding]=optargs{:
};
%usememorablevariablenames
m=window
（1）;
n=window
（2）;
if~mod（m,2）m=m-1;
end%checkforevenwindowsizes
if~mod（n,2）n=n-1;
end
if（ndims（image）~=2）%checkforcolorpictures
display（'
Theinputimagemustbeatwodimensionalarray.'
）
Considerusingrgb2grayorsimilarfunction.'
return
%Initialization.
[rowscolumns]=size（image）;
%sizeoftheimage
%Padtheimage.
imageP=padarray（image,[（m+1）/2（n+1）/2],padding,'
pre'
imagePP=padarray（imageP,[（m-1）/2（n-1）/2],padding,'
post'
%Alwaysusedoublebecauseuint8wouldbetoosmall.
imageD=double（imagePP）;
%Matrix'
t'
isthesumofnumbersontheleftandabovethecurrentcell.
t=cumsum（cumsum（imageD）,2）;
%Calculatethemeanvaluesfromthelookuptable'
.
imageI=t（1+m:
rows+m,1+n:
columns+n）+t（1:
rows,1:
columns）...
-t（1+m:
rows+m,1:
columns）-t（1:
rows,1+n:
columns+n）;
%Noweachpixelcontainssumofthewindow.Butwewanttheaveragevalue.
imageI=imageI/（m*n）;
%Returnmatrixintheoriginaltypeclass.
image=cast（imageI,class（image））;
%BERNSENlocalthresholding.
%BW=BERNSEN（IMAGE）performslocalthresholdingofatwo-dimensional
%arrayIMAGEwithBernsen'
sthresholding.Themethodusesa
%user-providedcontrastthreshold.Ifthelocalcontrast（max-min）is
%aboveorequaltothecontrastthreshold,thethresholdissetatthe
%localmidgreyvalue（themeanoftheminimumandmaximumgreyvaluesin
%thelocalwindow）.Ifthelocalcontrastisbelowthecontrast
%thresholdtheneighbourhoodisconsideredtoconsistonlyofoneclass
%andthepixelissettoobjectorbackgrounddependingonthevalueof
%themidgrey.
%BW=BERNSEN（IMAGE,[MN],CONTRAST_THRESHOLD,PADDING）performslocal
%thresholdingwithoddvaluedM-by-Nneighbourhood（defaultis3-by-3）.
%ThedefaultvalueforCONTRAST_THRESHOLDis15.Todealwithborder
%pixelstheimageispaddedwithoneofPADARRAYoptions（defaultis
）.
%imshow（bernsen（imread（'
）））;
%SeealsoPADARRAY,RGB2GRAY.
1.0$$Date:
2013/03/0916:
functionoutput=bernsen（image,varargin）
%Initialization
%onlywant3optionalinputsatmost
3
PossibleBERNSENparametersare:
（image,[mn],contrast,padding）'
optargs={[33]15'
%setdefaults
[window,contrast_threshold,padding]=optargs{:
ifndims（image）~=2
Theinputimagemustbeatwo-dimensionalarray.'
ifsum（mod（window,2））~=2
Sorry,onlyoddvaluedwindowdimensionsaresupported'
%Converttodouble
image=double（image）;
%Meanvalue
mean=averagefilter（image,window,padding）;
%Localcontrast
local_contrast=maxfilt2（image,window）-minfilt2（image,window）;
%Initializetheoutput
output=zeros（size（image））;
%Whenevercontrastinthewindowislowassumehomogenousarea
mask=local_contrast<
contrast_threshold;
output（mask&
image>
=128）=1;
%Otherwisecomparetothemeanvalue
output（~mask）=（image（~mask）>
=mean（~mask））;
functionY=maxfilt2（X,varargin）
%MAXFILT2Two-dimensionalmaxfilter
%Y=MAXFILT2（X,[MN]）performstwo-dimensionalmaximum
%filteringontheimageXusinganM-by-Nwindow.Theresult
%YcontainsthemaximunvalueintheM-by-Nneighborhoodaround
%eachpixelintheoriginalimage.
%ThisfunctionusesthevanHerkalgorithmformaxfilters.
%Y=MAXFILT2（X,M）isthesameasY=MAXFILT2（X,[MM]）
%Y=MAXFILT2（X）usesa3-by-3neighborhood.
%Y=MAXFILT2（...,'
shape'
）returnsasubsectionofthe2D
%filteringspecifiedby'
:
full'
-Returnsthefullfilteringresult,
same'
-（default）Returnsthecentralfilterareathatisthe
%samesizeasX,
valid'
-Returnsonlytheareawherenofilterelementsareoutside
%theimage.
%Seealso:
MINFILT2,VANHERK
[S,shape]=parse_inputs（varargin{:
}）;
%filtering
Y=vanherk（X,S
（1）,'
max'
shape）;
Y=vanherk（Y,S
（2）,'
'
col'
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function[S,shape]=parse_inputs（varargin）
shape='
;
flag=[00];
%sizeshape
fori=1:
nargin
t=varargin{i};
ifstrcmp（t,'
）&
flag
（2）==0
shape='
flag
（2）=1;
elseifstrcmp（t,'
elseifflag
（1）==0
S=t;
flag
（1）=1;
else
error（['
Toomany/Unkownparameter:
t]）
ifflag
（1）==0
S=[33];
iflength（S）==1;
S
（2）=S
（1）;
iflength（S）~=2
Wrongwindowsizeparameter.'
functionY=minfilt2（X,varargin）
%MINFILT2Two-dimensionalminfilter
%Y=MINFILT2（X,[MN]）performstwo-dimensionalminimum
%YcontainstheminimunvalueintheM-by-Nneighborhoodaround
%ThisfunctionusesthevanHerkalgorithmforminfilters.
%Y=MINFILT2（X,M）isthesameasY=MINFILT2（X,[MM]）
%Y=MINFILT2（X）usesa3-by-3neighborhood.
%Y=MINFILT2（...,'
MAXFILT2,VANHERK
min'
functionY=vanherk（X,N,TYPE,varargin）
%VANHERKFastmax/min1Dfilter
%Y=VANHERK（X,N,TYPE）performsthe1Dmax/minfilteringoftherow
%vectorXusingaN-lengthfilter.
%ThefilteringtypeisdefinedbyTYPE='
or'
.Thisfunction
%usesthevanHerkalgorithmformin/maxfiltersthatdemandsonly3
%min/maxcalculationsperelement,independentlyofthefiltersize.
%IfXisa2Dmatrix,eachrowwillbefilteredseparately.
%Y=VANHERK（...,'
）performsthefilteringonthecolumnsofX.
）returnsthesubsetofthefilteringspecified
%by'
%Xcanbeuint8ordouble.IfXisuint8theprocessingisquitefaster,so
%dont'
tuseXasdouble,unle
升级会员 