[gdal-dev] How to make my GDAL code faster?

[Nicolas DEGARNE]

Hello,


Just to introduce, I am a student in internship in French Guyana.I'm French
and I have not wrote in English thing's for a wile so sorry for my bad
English.

I'm a beginning in GDAL in c (for the moment) with Ubuntu

I am trying to make a program extracting target from pictures in
tif(geotiff) for example Size X: 17553 // Size Y: 15226 ..

To detect target,I used a algorithm which get the mean of a windows 5x5 and
compare the energies off the windows with the central pixel.

I wants to improve my software in order to be faster  because for the moment
it spend 200s

To run through every pixel I used the process below

        hScanBloc1   = (GByte *) CPLMalloc(sizeof(GByte)*5);
        hScanBlocOut = (GByte *) CPLMalloc(sizeof(GByte)*1);


for( iLine = 0; iLine < nYSize-5; iLine++)
>             {
>             for( iCol = 0; iCol < nXSize-5; iCol++ )
>                 {
>
>                 //Get line from the windows
>                 eErr1 = GDALRasterIO( hBand1,     GF_Read,
> iCol,      iLine  ,   5,   1, hScanBloc1,      5,1,  GDT_Byte, 1, 1);
>                 eErr2 = GDALRasterIO( hBand1,     GF_Read,        iCol,
> (iLine+1),   5,   1, hScanBloc2,      5,1,  GDT_Byte, 1, 1);
>                 eErr3 = GDALRasterIO( hBand1,     GF_Read,        iCol,
> (iLine+2),   5,   1, hScanBloc3,      5,1,  GDT_Byte, 1, 1);
>                 eErr4 = GDALRasterIO( hBand1,     GF_Read,        iCol,
> (iLine+3),   5,   1, hScanBloc4,      5,1,  GDT_Byte, 1, 1);
>                 eErr5 = GDALRasterIO( hBand1,     GF_Read,        iCol,
> (iLine+4),   5,   1, hScanBloc5,      5,1,  GDT_Byte, 1, 1);
>
>                 //Get all value in the     Windows
>                 for(iWindows=0;iWindows<5;iWindows++)
>                 {
>                     Value=Value+hScanBloc1[iWindows];
>                     Value=Value+hScanBloc2[iWindows];
>                     Value=Value+hScanBloc3[iWindows];
>                     Value=Value+hScanBloc4[iWindows];
>                     Value=Value+hScanBloc5[iWindows];
>                 }
>



>                                         hScanBlocOut[0]=FinalValue;
>                                         eErr = GDALRasterIO( hBandOut1,
> GF_Write,iCol,iLine ,1,1 , hScanBlocOut,   1,1,  GDT_Byte, 1, 1);
>
>
>
>                                       CPLErr eErr;
>                                       eErr = GDALRasterIO( hBandOut1,
> GF_Write,iCol,iLine ,1,1 , hScanBlocOut,   1,1,  GDT_Byte, 1, 1);
>


I read on help and *gdal.h* File Reference that it could exist faster way so
I have some ideas like :
             Merging the code to c++
             Using Read/Write block
             Use Tile/Block to accelerate the process
             Use the Warpprocess to calulate windows mean's faster

I hope you understand my request and it was clear

Thanks for advance

Best Regards

Nicolas

-- 
Nicolas Degarne
76 rue de Nancy
94170 Le Perreux
06 84 93 80 94
nicolas.degarne at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/gdal-dev/attachments/20090415/400b93e4/attachment-0001.html