Jump to content
PDS Geosciences Node Community

Ray Arvidson

  • Posts

  • Joined

  • Last visited

Everything posted by Ray Arvidson

  1. Hi: In your message "ERDAS it says there are no bands in the information". Not sure what you mean. ERDAS recognizes that it is an image cube but does not provide the wavelengths for each band? ERDAS does not know that it is a multiband cube, or what? Ray Arvidson
  2. Never flatten a map projected product. Flattening is only for sensor space products. Please read the Software Specification Document on the PDS site under CRISM to understand the nature of the various product types. For MTRDR nothing more than examining the spectra and associated products is required. For the TRDR IF products use CAT to do a volcano scan correction and map projection. Generate spectral parameters in sensor space, flatten them, and then project them. Then do your science. Ray
  3. I mean under data filtering ,choose flatten summary products. This is because the spectral summary products vary from column to column. Flattening does some column wise smoothing to make a better product. Do this in sensor space. Ray Arvidson
  4. Sehaj: Both of these apps are a out of date and are not needed with the TRR3 or MTRDR data sets. I suggest skipping their use. Do flatten your spectral parameter sensor space data before projecting. Ray Arvidson
  5. I am passing your question about map projection issues onto Feng Zhou, who is our staff member working with maps and ENVI CRISM data. I will also have Feng take a look at the destriping issue with the HSP data. Ray Arvidson
  6. Sehajpal: Destriping must be done on the sensor space data. That way each column is from one spectral detector set and destriping takes advantage of this simple arrangement. Did you try destriping after projection? You state: "Projecting the cube and applying destriping (Display 3 in screenshot) gives unusual results." If so, that is likely your problem. Ray Arvidson
  7. I just searched the Mars Orbital Data Explorer and there are hundreds of FRT observation from 2012 until now. Look for TRR3 TRDRs and FRTs and/or ATOs. Please read the CRISM Software Specification Document to become familiar with CRISM products https://pds-geosciences.wustl.edu/mro/mro-m-crism-2-edr-v1/mrocr_0001/document/crism_dpsis.pdf
  8. Hi: I would use the TRR3 data sets and the CRISM Analysis Tool that works with ENVI. Are you using ENVI? Ray Arvidson PDS Geosiences Node Manager CRISM Team Member
  9. HiRISE and CRISM map projected data will not exactly overlay because they have slightly different pointings and associated errors. In addition, sometimes one needs to use a GDAL app (see attachment) to first change the map information for HiRISE to be consistent with the CRISM projection. I employ the GDAL app to get them closer and then I use ENVI's Image Registration Workflow with lots of control points and a triangular warp to do fine registration of HiRISE to CRISM. An issue associated with the ENVI app is that the difference in resolution between the two data sets cannot be more than a factor of six (check this number, I may be wrong). Thus, first down sample the HiRISE image to be amenable to use with CRISM data coregistration. So, the HiRISE data should have 3 m/pixel to work with 18 m/pixel ENVI data. jue.txt
  10. The conversion is likely included in the code used to generate the BR at JHU/APL. I am sure various languages have this utility. Why do you need to know how it was done? Ray
  11. The spectral parameter maps from CAT are single band images in floating point. They are not RGB files. To colorize them I use ENVI's Change Color Table and switch to Rainbow. Other color palettes are also available. Not sure why you would want to loose information by converting them to byte RGB files. Ray
  12. Hi Nathan: Multiple questions here. First you can do math on a spectrum by choosing Spectral Math under the Spectral tool on the ENVI menu. This is not a CAT tool, rather an ENVI tool. Once you are happy with whatever you derive you can switch to ENVI's Band Math to apply your new spectral parameter math to a hyperspectral image, e.g. CRISM data. Then you will have a parameter value output image to work with, colorizing or whatever you wish to do. Make sure you save your Band Math equation. Again, these steps do not use CAT. Ray Arvidson CRISM Science Team Member and Geoscience Node Manager
  13. You need files of the same size to do New File Builder. Read up on making and using masks to subset another data set. Ray
  14. OK, now I see what you mean. These 1.2 micrometer features could be real absorptions. Smectite clays present such an absorption. Look at the ENVI spectral library that is part of CAT. Given that my job is to help you get to the spectra and interpret without interpreting artifacts, I will leave it to you to do the assignments. Ray
  15. You must join files of the same size, in terms of lines and samples. Subset the larger file using the smaller file to generate a mask of for the smaller size and then use the mask to subset the larger file to fit the size of the smaller file. Then use New File Builder. Read the instructions on doing New File Builder and you will see where you went astray. Rochdi, as before in our many emails, may I suggest that you need to read the documentation before embarking on a new process. Ray
  16. You lost me. Please resend the plots with the features labeled at 1.2 micrometers that you think might be due to mineralogy. Ray
  17. These are spike artifacts in the data. Sometimes the instrument individual detectors are ill behaved and this is an example. Sometimes it is associated with going over sharp albedo boundaries where a detector has memories of what it was just sampling. There are no minerals that I know of that have such narrow single detector absorptions. Please ignore these spikes. Ray
  18. Please attach a spectrum that shows this feature. I strongly suspect that it is an instrument artifact. Include the scene ID from which the spectrum was retrieved. Ray Arvidson PDS Geosciences Node Manager CRISM Science Team Member
  19. No need to run IDL directly. Just launch the New File Builder under Raster Management tool bar to combine the two cubes. Ray
  20. After separately registering the S and L mapped data to the CTX data I would use New File Builder under Raster Management to make an S, L combined mapped cube. Ray
  21. Because the bad bands are applied in CAT using the HDR record and in the spectral display. Remove them manually from the text file using the bad band list in the HDR list.
  22. Bad bands have been applied to TRR3 products. TRR2 products should not be used as they represent only an initial phase of calibration. You can see which bands have been assigned as bad by looking at the ENVI header records. Ray Arvidson
  23. The DDR has elevation data from MOLA observations. In sensor space so just project the DDR using CAT and use the projected MOLA elevation plane. Otherwise no special DEMs for CRISM data. Ray
  24. You cannot do a spectrum from a monochrome image. Overlay the spectral index map over the TRDR data. Then select the TRDR data on the ENVI vertical toolbar and then you can see the spectra for bright areas in the spectral index map, if you move the cursor the the bright pixels in the spectral index map. Ray
  25. Flatten is used to reduced column to column variations in sensor space TRDRs. The TERs have already been flattened. Ray
  • Create New...