Hyperspectral GeoPDF
Hyperspectral Imagery has been an interest of mine for some time now and I think I've found a pretty good way to display it in a map. I used Map2PDF for Raster and quickly added the first 54 bands (out of 224). I assigned each band an arbitrary color of my choosing for this example but a more qualified Remote Sensing Scientist could assign any color they want to as many bands as they want. There is NO LIMIT on the amount of bands a user can display. This image is NOT geo-referenced so please don't look for any spatial information in there but I would encourage you to download the data from here or from the Menu option in this flash viewer.
Again, this is AVIRIS data that I quickly converted to GeoPDF and uploaded to Acrobat.com so that anyone can download and view it. When you do download the entire file, open the Layers tab to gain access to all 54 bands. I think a qualified eye will be able to see the possibilities with this one! Enjoy...
Labels: AVIRIS, Hyperspectral, Map2PDF, Raster
posted by Adam Estrada at 4:14 PM
4 Comments:
The advantage of HyperSpectral Imagery (HSI) is in its use for geospatial feature extraction through imaging spectroscopy. How do you see the HSI data being used after its conversion to GeoPDF format? Thanks.
8:13 PM EDT
My understanding of imaging spectroscopy is that its the "simultaneous acquisition of spatially coregistered images in many spectrally contiguous bands". PDF allows the user to dynamically enable and disable each band as they see fit and in as many combinations as they want. The challenge for making this work in GeoPDF is whether or not we can add subsequent bands to the PDF and get useful information out of it. For this quick example, I added 54 bands and assigned a color to each one. I am going on the assumption that someone with knowledge of the AVIRIS sensor would know what the wave lengths are for each band and could combine them as needed to find the features they are looking for.
8:42 PM EDT
Does the GeoPDF maintain the original byte value of the HSI data layer at each pixel in the image?
Imaging spectroscopy is more than just collecting the imagery, it is about identifying and classifying features based on unique spectral signatures at each pixel in the image. If the conversion to GeoPDF removes the calibrated byte value in each band via compression, then the ability to generate classifications based on these images is greatly reduced.
4:44 PM EDT
Hi Paul,
Our raster product does not necessarily maintain the original byte values for the image. In Acrobat today there is no provision to get to the individual pixel per band from the image. What it does allow is the ability to add band-level data in various blend methods to be viewed at run time. The data would obviously have to be processed in something like IMAGINE or ENVI first and then exported to GeoPDF and as it stands today, the creation of each GeoPDF is entirely up to the creator. Again, this product is in beta right now so we are looking for feedback! What would make this application of GeoPDF more useful to someone looking to perform multispectral analysis?
Adam
5:33 PM EDT
Post a Comment
<< Home