Error module 'acolite.aerlut' has no attribute 'interp2d' for WorldView

[quinten]

Hi Thang

ACOLITE uses a single estimate of the atmospheric path reflectance (rho_path), so this shift in histogram between rho_t and rho_s is normal. The fact that you are getting negatives is because the rho_path it is overestimated, even for the Rayleigh only (use of minimum tau_a = 0.001) atmosphere. The overestimation could have a few origins, e.g. satellite geometry, LUT interpolation, sensor calibration, image format...

You can add a print(rhod) statement for example above the loop that you edited before, although it may not give us much further information in this case seeing as most of your image has rhos_428 < 0. It depends really on your intended use of this data whether you can use this data as it is now.

Quinten

[hnthang]

Hi Quinten,

We intend to use this WV2 scene for vegetation classification (11 classes). Since we are not looking for an absolute reflectance value, the image itself may be used somehow. I will compare this corrected scene with one corrected using FLAASH in ENVI from my college.

So, what do you think the main problem regrading our scene at the moment (since several causes were listed from your post) ? And can we improve this correction in next release of acolite_mr for WV?

Thang

[quinten]

Hi Thang

I am not sure, perhaps the comparison with FLAASH will teach us something.

Quinten