#194 Supervised classification with table definition not working



I try to do some work with remote sensing imagery. I try doing some supervised classification using Maximum Likelihood and firstly i do it with some training areas (a shapefile). At this point, all is well, the grid is correctly classified. It gives me at output a table with some statistics for each class on all bands. But when i try to use this table definition in a new supervised classification using table statistics from the previous training areas classification, it outputs a "blank" grid or a grid with the the same value for all the cells, even if i do it to the same grid previously tested.

Here attached my dbf table definition.

Thanks for any help.

1 Attachments


  • Hi,

    we did a quick test with your table in SAGA 2.1.2 and encountered no problems - our imagery was "classified" in the end. Do you use an older SAGA version? Please note that it is necessary that the order of the input grids matches the order of the columns in the table.

  • AlexMex

    I use 2.1.2 in 64 bits edition. Is the order for the input grids given by the order in the list or by the number in the left (see my attachment) ? This number (01, 02...), i see by testing, is given by the order when you load the grid in file menu (01 is the first grid loaded, 02 the second etc.).

    See in the second capture if my parameters are right for classification from table statistics.

    Last edit: AlexMex 2014-07-10
  • The order of the input grids is determined from the order of the grids in the grid list. The grid at the top of the list is the first grid and the grid at the bottom of the list is the last grid. In your screen captures you have the following order: MICHA_PIR, MICHA_R, MICHA_G, MICHA_B. This must be the same order which you used to create the statistics table.

  • AlexMex

    OK, the latest all run well. I don't remember how reproduce my problem, maybe using 'Normalize' parameter.



Cancel   Add attachments