hello,

I am trying to make a band math calculation by this formula
exp10(-0.29-0.708* log10(Rrs_488/ Rrs_547)+1.12* pow(log10(Rrs_488/ Rrs_547),2)) [1]

I use L2 MODIS file.

As a result I get a black image with only few grey pixels.. If is calculate the new band by only equation

-0.29-0.708* log10(Rrs_488/ Rrs_547)+1.12* pow(log10(Rrs_488/ Rrs_547),2) [2]

I get normal image result. But once i put it into () of exp10 (same with exp and pow functions) - i get this black one.
I checked the values of the image calculated by formula [2] - they are positive or negative in different areas.

I also tried to calculate the band by equation [2] and then use it's values in next calculations by exp10(new band) - but result is the same..

If take any pixel value from that range and calculate, for example, POWER(10,X) function in excel with that values on X - I get normal positive values.

Parts of formula works, but it all - not. What is wrong, what should I change? May be there are some limitations somewhere on writing expressions..?

Thank you very much
Evgenia

Hello Thomas

Thanks you for your reply
I checked the result - yes, pixel info shows different positive values, you are right. But I did not find a way to make the normal image. In Color manipulation and also in histogram i don't see any data range, any tries of adjustment were unsuccessful.. The data range is mostly between 0 and 1 (by pixel info) - but..
I can upload the file - but i have no FTP, so...

Evgenia

Hello Thomas

the values seem to be correct, and looks really like display problem.
But i still did not find a way to solve it..both histogram and color manipulation sections don't show any data..

Evgenia

Hi Evgenia,

ok, I had a look at the band computed by the equation you provided and I see what you mean.
Anyway, there is no error at all; only the colour manipulation does not handle the resulting ranges well, because there are some extreme outliers. If you 'mask out' these outliers, e.g. by creating an additional band by band maths, you will receive a perfectly valid picture.
I used the expression new_band_1 < 1 ? new_band_1 : NaN to replace every value greater than 1 by NaN. new_band_1 of course is the band computed by your expression.

I hope this helps.

Best regards,
Thomas

Dear Evgenia,



You can't use "<" and ">" directly in an XML file. You have to write "&lt;" instead of "<" and "&gt;" instead of ">". Sorry for this inconvenience (I hate it as well). Alternatively, you can encode the entire expression as CDATA. For example:



See http://www.w3schools.com/xml/xml_cdata.asp

Best regards
Norman

Hi Evgenia,

Sorry for not catching up for so long. Do you still suffer from your problem? If so, please attach the XML file, there must be something wrong with the file format, since "CDOM_Kow" is not part of your expression.

Norman

Hi Evgenia,

it seems you have confused the tags 'valid expression' and 'expression' in some cases.
The expression determines the target values of the band, and each band you reference in the expression must exist in the source product. This seems to be wrong for the target band 'CDOM_Kow', where you've entered the expression 'CDOM_Kow' -- which does presumably not exist in the source product, so the error is thrown. Please check the other target bands you've specified, too.
The valid expression determines the constraints the pixels have to fulfill if they are to be considered. You have done this right in some cases (e.g. 'not landMask' seems sensible for me), but in the 'CDOM_Kow' case I think you have entered the expression as valid expression.

Hope this helps,
best regards,
Thomas