| FFT on time series NDVI image [message #89821] |
Fri, 05 December 2014 01:22 |
prasenjit acharya
Messages: 2
Registered: December 2014
|
Junior Member |
|
|
Hi all,
is anybody knows about the Fast Fourier Transformation (FFT) on multiband (stacked) time series NDVI image? I am a new user of IDL. I have borrowed some code over internet. but it returns some unusual results.
;start envi session
envi
;get the image
envi_select, title='Choose multispectral image', $
fid=fid, dims=dims,pos=pos
num_cols = dims[2]-dims[1]+1
num_rows = dims[4]-dims[3]+1
num_bands = n_elements(pos)
; BIL array
image = fltarr(num_bands,num_cols,num_rows)
for i=0,num_bands-1 do image[i,*,*] = $
envi_get_data(fid=fid,dims=dims,pos=pos[i])
;create empty data matrix to store the transformed image
image2 = fltarr(num_bands,num_cols,num_rows)
;fft on image
for S=0,num_bands-1 do image2[*,S] = FFT(image[*,S])
real = FLOAT(image2)
imagin = ATAN(IMAGINARY(image2), FLOAT(image2))
; the data back in BSQ format
image3 = fltarr(num_bands,num_cols,num_rows)
for i = 0, num_bands-1 do image3[*,*,i] = $
reform(real[i,*],num_cols,num_rows)
; get the map info of the original data
map_info = ENVI_GET_MAP_INFO(fid=fid)
;set output directory
output_dir = 'H:\'
filename = output_dir+'FFTR1'+'.tif'
;write file
ENVI_WRITE_ENVI_FILE, real, out_name=filename,map_info=map_info
end
any suggestion to modify the code is welcome.
thank you
Prasenjit
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
