A executable application that allows the user to open CRISM (either IR(L) or MTRDR) images and label spectra as interesting/mineral spectra, i.e. containing information on the surface (in the form of recognizable mineral absorption features) or uninteresting spectra, for e.g. dust or containing only continuum, noise and artifacts
. Installers are available for:
Windows
Linux
Mac
The Project
The main-focus of this project is to use supervised techniques to identify the various mineral spectral present in the CRISM image database. As the first step of this process we attempt to automate the decision on whether a given spectra in the CRISM database contains useful surface information or is simply an uninteresting spectra, for e.g. composed of dust, or contains only noise and artifacts.
The App
The app is simple standalone app designed in MATLAB ® that allows experts to open any CRISM IR images either (MTRDR or IR(L)) in the CAT format. The app opens the image and displays two Spectra – RED and GREEN. The app also allows the expert to move the position of the RED and GREEN spectra and label them appropriately.
INSTALLATION
Once you download the appropriate installation package for your OS. Place the installation package in an empty folder. Following this:
1) For Windows: Run the executable installer and follow along with prompts that will help you install the app and the MATLAB ® Compiler Runtime (MCR). Following this you can run the executable CISLapp_v1created in this folder by double-clicking on it.
2) For Linux: Unzip the package in the empty folder:
(a) Now also unzip the 'MCRInstaller.zip file '
(b) Now in the terminal navigate to the folder where you have unzipped the 'MCRInstaller.zip file' and type in ./install. This will lead to a series of prompts that will help you install the MCR.
(c) Once MCR is installed you run the app from the terminal as ./run_CISLapp.sh MCRROOT/v717 where MCRROOT is the location where you installed MCR.
3) For Mac:
(a) Run the installer by double-clicking the downloaded 'MyAppInstaller_mcr.app' file
and follow the prompts to install the app and the Matlab Compiler Runtime (MCR). Keep track of the folder where you install the app (APPFOLDER) and the folder in which you place the MATLAB Compiler runtime (MCRROOT),
(b) Navigate to the folder with the shell script to run the app as cd APPFOLDER/application/ .
(c) You run the app from the terminal as ./run_CISLapp_mac_v1.sh MCRROOT/v83 .
Using the App
To use the app:
1) First enter the user-name as prompted (on the top left of the app). All label files are saved with the user-name please use the same name if possible.
2) Then click the button 'Open Image' to get an address explorer. Use this to open CRISM images of the IR (L) or Joint (J) type. The app is set up to use only CAT format images. This is the SLOWEST part, please wait a few seconds for the image to load.
3) Once you provide the image address the image will plot two spectra on the plot screen. On is a Numerator Spectra (plotted in green) and a Denominator (in red). The user can control the position of the numerator and denominator spectra using the various methods of browsing the image as shown below.
4) To move around in the image the user can choose the following options:
a) Use the mouse to click on the region in the of the image you want to navigate to.
b) Use the arrow keys to move one spectra to the left or the right.
c) Or enter the row or column you want to go to in the appropriate textbox.
5) Once you find a spectrum you want to label. Click the appropriate button. The buttons indicate:
a) "Mineral/Interesting": The spectra show features which indicate surface mineral properties.
b) "Neutral/Uninteresting": The spectra contains only noise and atmospheric distortions.
c) "Noisy/Not Clear": Not sure either way or too noisy to tell.
6) When you close the App, the App will automatically attempt to send a copy of the labels to a FTP server using passive FTP. If this fails for some reason (not connected to internet, server restart etc..), the app will prompt you to simply e-mail the labels and spectra to This email address is being protected from spambots. You need JavaScript enabled to view it.. The app also creates a local copy of the label, these files are saved as .csv files in your App Folder.
If there are any questions or comments please contact the System Admin (Arun Saranathan). Thank you for your help and patience.