双三次插值python实现_双三次插值Python.pdfVIP

双三次插值python实现_双三次插值Python

我⽤Python编程语⾔开发了双三次插值来演⽰给⼀些本科⽣。

⽅法如wikipedia所述，

代码运⾏良好，只是得到的结果与使⽤scipy库时得到的结果略有不同。

插值代码在下⾯的函数bicubic_interpolation中显⽰。importnumpyasnp

importmatplotlib.pyplotasplt

frommpl_toolkitsimportmplot3d

fromscipyimportinterpolate

importsympyassyp

importpandasaspd

pd.options.display.max_colwidth=200

%matplotlibinline

defbicubic_interpolation(xi,yi,zi,xnew,ynew):

#checksorting#checksorting

np.any(np.diff(xnew)0)andnp.any(np.diff(ynew)0):

raiseValueError(dataarenotsorted)

ifzi.shape!=(xi.size,yi.size):

raiseValueError(ziisnotsetproperlyusenp.meshgrid(xi,yi))

z=np.zeros((xnew.size,ynew.size))

deltax=xi[1]-xi[0]

deltay=yi[1]-yi[0]

forn,xinenumerate(xnew):

form,yinenumerate(ynew):

ifxi.min()=x=xi.max()andyi.min()=y=yi.max():

i=np.searchsorted(xi,x)-1

j=np.searchsorted(yi,y)-1

x0=xi[i-1]

x1=xi[i]

x2=xi[i+1]

x3=x1+2*deltax

y0=yi[j-1]

y1=yi[j]

y2=yi[j+1]

y3=y1+2*deltay

px=(x-x1)/(x2-x1)

py=(y-y1)/(y2-y1)

f00=zi[i-1,j-1]#row0col0x0,y0

f01=zi[i-1,j]#row0col1x1,y0

f02=zi[i-1,j+1]#row0col2x2,y0

f10=zi[i,j-1]#row1col0x0,y1

f11=p00=zi[i,j]#row1col1x1,y1

f12=p01=zi[i,j+1]#row1col2x2,y1

f20=zi[i+1,j-1]#row2col0x0,y2

f21=p10=zi[i+1,j]#row2col1x1,y2

f22=p11=zi[i+1,j+1]#row2col2x2,y2

if0ixi.size-2and0jyi.size-2:

f03=zi[i-1,j+2]#row0col3x3,y0

f13=zi[i,j+2]#row1col3x3,y1

f23=zi[i+1,j+2]#row2col3x3,y2

f30=zi[i+2,j-1]#row3col0x0,y3

f31=zi[i+2,j]#row3col1x1,y3

f32=zi[i+2,j+1]#row3col2x2,y3

f33=zi[i+2,j+2]#row3col3x3,y3

elifi=0:

f03=f02#row0col3x3,y0

f13=f12#row1col3x3,y1

f23=f22#row2col3x3,y2

f30=zi[i+2,j-1]#row3col0x0,y3

f31=zi[i+2,j]#row3col1x1,y3

f32=zi[i+2,j+1]#row3col2x2,y3

f33=f32#row3col3x3,y3

elifj=0: