Biophysical Processor for external image data

The “Biophysical Processor” found in the Thematic Land Processors menu can easily be run with S2-data. Now I am trying to get my parameters from an external image. It is an ENVI image of a field campaign we had last summer. SNAP allows me to open the ENVI-image and display its content, but I can’t run the Biophysical Processor. Precisely it says:

“Product XXX.bsq does not contain a raster with name ‘view_zenith_mean’”

This makes perfect sense, since the geometry of sun and sensor need to be specified. And I do have this information, also as ENVI-images. But how do I include these into the processor? How do I supply the desired geometry rasters?

Hi,

you can try adding a new band in the product XXX.bsq with the name “view_zenith_mean” with the corresponding data. You can do it with the BandMaths operator ( http://forum.step.esa.int/t/create-a-new-band-in-a-beam-dimap-product-from-another-products-band/).

1 Like

Thank you, obarrilero, that was an excellent idea! I added band by band and named them according to what the error messages asked me for (“view_zenith_mean”, “sun_zenith”, “sun_azimuth”, “view_azimuth_mean”).

After that, I get:
“A problem occurred during processing the raget product processing.
Type: OperatorException
Message: java.io.EOFexception”

Looks like the processor is still missing further information, but now it won’t tell me what it is. Maybe the processor is only meant to be run with S2-data?

Hi,

Can you share the product (already pre-processed and containing all the bands) so that I can test it and debug the error?

Dear @offeltoffel,

did you try to convert your product to BEAM-DIMAP format before processing it? The error you got sounds like something related to the ENVI file format.

I could successfully use the biophysical processor to estimate parameters vegetation biophycical variables estimates from aerial hyperspectral data, supplied together with the “view_zenith_mean”, “sun_zenith”, “sun_azimuth”, “view_azimuth_mean” bands.

Since the method have been implemented to generate biophysical products from VEGETATION, MERIS, SPOT, and LANDSAT sensors, I tried to use those Remote Sensing datasets to estimate vegetation biophycical variables, and I got the following errors:

PROBA-VEGETATION: Missing band at 740.0 nm
MERIS: Missing band at 1610.0 nm
SPOT: Missing band at 865.0 nm
LANDSAT8: Missing band at 740.0 nm

I am wondering if the processor can not process those Remote Sensing datasets, and if the support for those sensors is planned to be implemented in the near future.

Another issue I would like to address is if the processor needs for geometry information bands (“view_zenith_mean”, “sun_zenith”, “sun_azimuth”, “view_azimuth_mean”) also for the retrieval of variables not directly dependent on the geometry of illumination or canopy structure (eg. FVC). If this is not the case, how to deal with the fact that some of the standard distribution format of Remote Sensing datasets do not contain any information about acquisition geometry (es. LANDSAT)?

Best,

Federico

1 Like