Interpolation of satellite image with overlapping areas

Question

I have a problem with satellite image interpolation. I understand how a satellite receives an image of the earth:

and I'm trying to interpolate using the following code:

import scipy as sc import scipy.interpolate as spi import numpy as np def interpolate_data(img, lat, lon): p_x, p_y = sc.mgrid[-1:1:lat.shape[0]*1j, -1:1:lon.shape[1]*1j] n_x, n_y = sc.mgrid[-1:1:img.shape[0]*1j, -1:1:img.shape[1]*1j] n_lat = spi.griddata((p_x.ravel(), p_y.ravel()), lat.ravel(), (n_x, n_y), method='linear') n_lon = spi.griddata((p_x.ravel(), p_y.ravel()), lon.ravel(), (n_x, n_y), method='linear') print "n_lat:",n_lat.shape print "n_lon:",n_lon.shape print "img:",img.shape return n_lat, n_lon 

But if I interpolate in this way, I will have the wrong data.

In my case, the satellite will take 9 images of the earth. Between curve A and curve B is the area that was taken by the satellite during the first shot. Between curve C and curve D is the area occupied by the satellite in the second image. E and F - area overlapping.

If I use the code above for interpolation, I will double count the overlapping area and give the wrong data. So I want to ask if there is a way to count these overlapping areas only once. Thanks.

If I use the code above for interpolation, I will double count the overlapping area and give the wrong data. So I want to ask if there is a way to count these overlapping areas only once. Thanks.

Answer 1

If your images have attached geographic data (for example, GeoTIFF files), you can consider using the GDAL library , this library is written in C / C ++, but also has a python package .

It can work with dozens of different formats and has several different image processing methods for converting and converting images.

Answer 2

Can't you just combine the data from all 9 pictures and interpolate all at once? I see no reason to consider these nine data sets separately. And if you have it, I think you better create 9 different images.