X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 09 Nov 2011 08:49 PM by  anon
Worldview-2 Atmospheric Correction with FLAASH - Negative Reflectance Values
 2 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
09 Nov 2011 08:49 PM
    Hi, I am currently attempting to atmospherically correct Worldview-2 imagery with FLAASH. However, after atmospheric correction I have many negative reflectance values. Original imagery statistics Basic Stats Min Max Mean Stdev Band 1 0 549 231.793323 19.082733 Band 2 0 856 241.299560 26.096208 Band 3 0 1406 270.464050 40.445499 Band 4 0 1122 188.705128 34.371135 Band 5 0 875 117.064133 27.352519 Band 6 0 1101 212.362145 83.927534 Band 7 0 1077 209.802802 114.767699 Band 8 0 1020 221.005078 130.223998 I've used the Worldview radiance calibration tool to convert to radiance. I specified to output to floating point and output radiance units are microWcm-2nm-1sr-1. Resulting calibrated radiance image statistics Basic Stats Min Max Mean Stdev Band 1 0.000000 10.789248 4.555329 0.375024 Band 2 0.000000 15.367094 4.331861 0.468485 Band 3 0.000000 16.830702 3.237624 0.484158 Band 4 0.000000 15.303263 2.573800 0.468797 Band 5 0.000000 16.823523 2.250778 0.525904 Band 6 0.000000 12.717864 2.453036 0.969463 Band 7 0.000000 13.333239 2.597354 1.420822 Band 8 0.000000 9.260119 2.006405 1.182245 mlzoffoli (Re: FLAASH Scale Factor and Radiance Units post) said to convert with band math to get radiance values to use this equation: (float(b1)*abscalfactor_b1/effectiveBandwidth_b1)/10. I have calculated radiance using this equation and the values are the same as when using the ENVI Worldview radiance calibration tool. Resulting radiance image statistics Basic Stats Min Max Mean Stdev Band 1 0.000000 10.789248 4.555329 0.375024 Band 2 0.000000 15.367094 4.331861 0.468485 Band 3 0.000000 16.830702 3.237624 0.484158 Band 4 0.000000 15.303263 2.573800 0.468797 Band 5 0.000000 16.823523 2.250778 0.525904 Band 6 0.000000 12.717864 2.453036 0.969463 Band 7 0.000000 13.333239 2.597354 1.420822 Band 8 0.000000 9.260119 2.006405 1.182245 Image statistics remain the same after conversion to BIL format. I used the radiance image as input to FLAASH. After running FLAASH, I have many negative reflectance values. FLAASH atmospherically corrected image statistics Basic Stats Min Max Mean Stdev Band 1 0 4027 1013.247076 116.350327 Band 2 -1180 6757 -608.006902 222.240073 Band 3 -1013 5986 -461.359516 191.361668 Band 4 -569 5899 6.609446 190.760860 Band 5 -372 7165 251.820297 227.588022 Band 6 -308 5390 511.701797 430.093464 Band 7 -348 7045 937.113198 783.974714 Band 8 -227 5346 888.287749 708.970879 I'm not sure how to resolve this issue. Any suggestions? Thank you.

    Deleted User



    New Member


    Posts:
    New Member


    --
    07 Jun 2015 03:50 AM
    I am facing the same problem. Did you find the problem?

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    08 Jun 2015 06:02 AM
    It is not at all unusual for negative values to be produced when using FLAASH. When you say 'many' can you determine which pixels are negative (perhaps make an ROI or mask of pixel values less than 0)? Are the negative pixels in areas of water or shadow (low radiance pixels)? If so, then it is sometimes the case that these pixels do not model well and can return negative reflectance, particularly in the blue bands. The negative values should not be the majority of the pixels in areas of reasonable radiance return.
    You are not authorized to post a reply.

    Forums Products ENVI