在Python中获取2D数组的索引

Question

I use osgeo from gdal module for processing geographic raster data.

#Get size of input raster (cols, rows)
cols = ds.RasterXSize #3531
rows = ds.RasterYSize #3314

So to read all the data, I create an array:

data = band.ReadAsArray(0, 0, cols, rows).astype(float)

Result of ReadAsArray is a 2D numpy array. Until here it works.

So in order to further apply the geographic transformation for each pixel in that raster, I need the column index and the row index for each pixel. I think I might need a numpy function to read that out, but I have no clue how that works with a 2D array.

I can access items the array by simply calling its index (eg band[0][1]), but I'd need the whole column index and the whole row index separately stored in eg col_idx and row_idx.

I tried with something like that, but didn't work:

for idx, val in enumerate(ints):
    print idx, val

Any help appreciated!

Answer 1

you can use numpy.meshgrid

i, j = numpy.meshgrid(range(ncols), range(nrows))
result = f(data, i, j)

Example

In [28]: nrows = 2 ; ncols = 4

In [29]: icol, irow = np.meshgrid(range(ncols),range(nrows))

In [30]: def a(r,c): return 100*r+c

In [31]: a(irow, icol)
Out[31]: 
array([[  0,   1,   2,   3],
       [100, 101, 102, 103]])

In [32]: 

data is XXXL?

In this case, instead of generating two large matrices with the row and column indices, you can use a for loop on the rows of your array (it's slower but not as slow as a double for loop)

icol = range(ncols)
for irow, row in enumerate(data):
    if irow>0:
         result = np.vstack((result, f(row, irow, icol)))
    else:
         result = f(row, irow, icol)