Professional Documents
Culture Documents
0 Users Guide
Users Guide
Version 4.x
October 2013
WARNING:
All information in this document, as well as the software to which it
pertains, is proprietary information of Applied Analysis Inc. and is subject
to an Applied Analysis Inc. license and non-disclosure agreement.
Neither the software nor the documentation may be reproduced in any
manner without the prior written permission of Applied Analysis Inc.
TRADEMARKS
ERDAS IMAGINE is a registered trademark of Intergraph Corporation.
This product includes GDAL (open source) software developed by The XFree86 Project,
Inc (http://www.xfree86.org/) and its contributors. Its use is subject to the following
conditions:
Copyright (c) 2000, Frank Warmerdam. Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal in the Software without restriction,
including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the
Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT
OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
OTHER DEALINGS IN THE SOFTWARE.
Other brands and product names are trademarks of their respective companies.
Table of Contents
1.1
Introduction to AAIC Image Calibration Software ...................................................................... 5
1.1.1
Background ........................................................................................................................... 5
2.1
Algorithm Description .................................................................................................................... 5
2.1.1
Performance Charateristics ................................................................................................... 7
2.1.2
Important Operational Notes ................................................................................................. 8
3.1
AAIC Software Operational Steps ............................................................................................. 8
3.1.1
Getting Started with the Software ......................................................................................... 8
3.1.2
Guidelines for Data Entry ...................................................................................................... 9
3.1.3
Step-by-Step Guide to Using the AAIC Software .............................................................. 9
3.1.3.1
Starting a Session ............................................................................................................. 9
3.1.3.2
Processing Setup ............................................................................................................ 11
3.1.3.3
AAIC Processing.......................................................................................................... 22
3.1.3.4
AAIC Results ............................................................................................................... 24
3.1.3.5
Expected Results from AAIC ....................................................................................... 25
Table of Figures
Figure 1. AAIC Processing Flow .................................................................................................................. 6
Figure 2. Selecting the Autonomous Atmospheric Correction Tool (AAIC) ........................................... 10
Figure 3. AAIC main menu window ......................................................................................................... 10
Figure 4. AAIC main menu window image file selection ...................................................................... 11
Figure 5. AAIC Main menu window Sensor selection .......................................................................... 12
Figure 6. AAIC Main menu window sensor type selected .................................................................... 13
Figure 7. AAIC Main menu window selecting new output location ...................................................... 14
Figure 8. AAIC Main menu window selecting output Material Id file .................................................... 15
Figure 9. AAIC Main menu window specifying output Material Id file .................................................. 16
Figure 10. AAIC Main menu window ready to launch processing ........................................................ 17
Figure 11. AAIC Main menu window selecting the Other sensor option .............................................. 18
Figure 12. AAIC Main menu window Other sensor option selected ..................................................... 19
Figure 13. AAIC Main menu window selecting the band center and widths file................................... 20
Figure 14. AAIC Main menu window band center and widths file selected ......................................... 21
Figure 15. AAIC Main menu window ready to run calibration using Other sensor type ....................... 22
Figure 16. AAIC Pre-processing status window ...................................................................................... 23
Figure 17. AAIC Reflectance image creation status window .................................................................. 23
Figure 18. IMAGINE Viewer with the results of AAIC processing ........................................................... 24
Figure 19. AAIC reflectance image swipe uncalibrated image vs. calibrated image. .......................... 25
Figure 20. Spectral plots for pixels covering vegetation.. ........................................................................... 26
Figure 21. AAIC reflectance image is on the left and the Material Id layer is on the right. ..................... 27
Figure 22. AAIC reflectance image is on the left and the Material Id layer is on the right. ..................... 28
Figure 23. AAIC reflectance image is on the left and the Material Id layer is on the right. ..................... 28
Figure 24. Poor calibration results, AAIC reflectance image on left and Material Id layer on right. ........ 29
Figure 25. Poor calibration results, AAIC reflectance image on left and Material Id layer on right. ........ 29
-Notes-
probe signature is associated with the reflectance spectrum of its detections in the
scene of interest, thereby allowing the correction factors for the scene of interest to be
determined. Figure 1 provides a summary flowchart of the AAIC process.
Multispectral or
Hyperspectral Image
Pre-Processing
(Creating Background Spectra)
Probe Signature
Processing
(Compare Backgrounds with
Probe Signatures, Find Best Fit)
Review AAIC
Results
The final stage of automated iteration enables the needed level of accuracy to be
achieved. This step of the processing uses a set of universal probe signatures derived
from reference scenes. Probe signatures are built into the software and the user does
not have to specify them. The software automatically compensates for sensor
differences or probe signature variations and, therefore, does not require the user to
select bright and dark materials within the scene or know which materials are present in
the scene or their location. By using specific land cover probe signatures, the process
automatically avoids clouds, high elevation shadows, and other artifacts, thereby
eliminating the need for operator interaction. The user only needs to specify the image
sensor attributes (spectral band centers and widths) and AAIC runs fully
autonomously. AAIC accommodates a wide variety of sensor types allowing for
calibration of imagery from just about any multispectral or hyperspectral sensor.
During the course of comparing calibrated image samples to probe signature spectra,
the image calibration process keeps track of which probe signatures perform best for
these samples and computes the band-wise ratio of the calibrated spectrum to the
probe spectrum. These ratios can be combined into a single figure of merit and
expressed as a percentage the confidence value. Images with high confidence values
contain good matches to the probe signatures and should be calibrated accurately.
Thus, the reported confidence value is a measure of how well the process performed.
10% land cover, for example, may cause reported pixel reflectance spectra to have
lower-than-expected reflectance accuracies, while reporting a Calibration
Confidence > 85%.
Zoom
Swipe
Pan (for moving around the image)
For a detailed discussion of these functions, the user should refer to the
documentation that is supplied with the ERDAS IMAGINE software.
3) From the main menu bar, click the Raster tab >
Autonomous Atmospheric Correction.
Radiometric >
Figure 2. Selecting the Autonomous Atmospheric Correction Tool from the ERDAS
IMAGINE tool bar
3) If the user has the image (to process) displayed in the Viewer, AAIC will
automatically fill-in the input image file name.
4) The software will pair down the sensor selection list based on the number
of bands in the image (i.e. 4 band image list excludes Landsat TM, since
images from this sensor typically contain at least 6 bands, not including
the thermal band)
5) The user should confirm that the correct sensor type is selected. Incorrect
sensor information will result in poor quality calibration results.
The current set of sensors restricts the use of AAIC to data from the
IKONOS, GEOEYE1, QuickBird2, Landsat MSS/TM, WorldView2, PLEIADES
and SPOT satellites. Additional sensors can be added upon request.
Important note: If the user wants to process an image from a sensor not
listed, there is an option to choose Other from the dropdown list and provide
a band center and width file. Details on using a *.bcw file can be found by
contacting AAI Customer support (see the end of this document).
When using SPOT or PLEIADES imagery, be sure that spectral bands are
properly ordered in the input image in terms of wavelength ascending order
(i.e. band 1 = blue band, band 2 = green band, , band 4 = NIR band).
GeoTIFF and JPEG2000 images for SPOT and PLEIADES typically have
their bands ordered such that band 1 = red, band 2 = green, band 3 = blue,
and band 4 = NIR. In their native format, the user would assign band 1 to red,
band 2 to green, and band 3 to blue to get a true color display of the image.
6) Once the sensor type has been selected, the user will now see that the
Reflectance Image File input window will be auto-populated with the root
image file name and a *_reflec added to the file name prefix. The output
file format is always IMAGINE img (except for GeoTIFF input, where the
reflectance image is output in GeoTIFF format), even if the input image is
in NITF format.
7) The user can redirect the output by clicking on the file browse button and
navigating to a new folder destination.
10) With all the processing options and output now selected, the user is ready
to launch the calibration process by clicking on the OK button.
11) If the sensor type is not provided in the drop down list, the calibration
process can be run by providing an external file which contains the sensor
band centers and widths.
a. After the user has provided the input image file name, clicking on
the Other option, found at the bottom of the drop down, will
prompt the user to provide a band center and widths file.
Figure 11. AAIC Main menu window selecting the Other sensor option
b. The user will now see a new input box in the Wizard window that
is used for specifying the band center and width file
Figure 12. AAIC Main menu window Other sensor option selected
c. Navigate to and choose the input band center and widths file.
Figure 13. AAIC Main menu window selecting the band center and widths file
d. If the user doesnt already have a *.bcw file, they will need to
supply Applied Analysis Inc. with specific details about the sensor
band passes. Specify the center wavelength and width (in the
same units as the center wavelength - nanometers) for each of
the spectral bands that comprise the sensor. For example, a 4
band IKONOS image will have 4 rows of entries, the first column
containing the center wavelength and a second column, which
contains the full band width.
Figure 14. AAIC Main menu window band center and widths file selected
12) The Batch button allows the user to accumulate processing runs into a
single batch run for immediate hands-off processing or scheduling for
launch at a later time. It utilizes the batch capability built into IMAGINE
(see Batch button highlighted in Figure 15).
Figure 15. AAIC Main menu window ready to run calibration using Other sensor type
1) As the AAIC process runs, various status windows will appear in the
IMAGINE Session Log window.
2) When AAIC has reached completion, the user will notice that the
session log will report a completed status. For a successful run, the user
should see the message that aicwrapper_aaic.exe exited normally. If
errors were encountered during the run, the user may notice an exiting
status other than successful. Scrolling back through the processing log will
reveal at what point in the processing the error occurred, and also the
probable cause.
3) After the image has been processed, the user can load the reflectance
image into the IMAGINE Viewer to compare it with the un-calibrated input
image.
The output reflectance image is the most important of the output files from the
AAIC process. This image contains pixels that have been converted to values
of material reflectance during the AAIC process. Images in this format, for
example, can be used with spectral matching algorithms to allow for the
identification of materials in the image using lab derived material spectra.
The image is output as a signed 16 bit image but can be scaled back to 8 bit
depth if desired, although some spectral detail may be lost by doing so.
Before Calibration
After Calibration
Figure 19. AAIC reflectance image swipe uncalibrated image on the left and calibrated
image on the right.
Figure 20. Spectral plots for pixels covering vegetation. The top plots shows the spectral
response before AAIC and the bottom plot shows the response after calibration. Note
the suppression of values in the blue band (band 1), since this band is most influenced by
haze in the atmosphere.
A few items to note that may indicate a problem with the input image and/or the
AAIC processing.
1) If the reported confidence falls below 70%, it is likely that AAIC did a
poor job with the matching of probe signatures and the calibration may be
in question. Confidence levels below 70% will not produce good results
using follow-on processing algorithms (i.e. spectral matching), and the
user should consider using another image if available. Confidence should
be in the range of 85% - 95% to assure a successful run.
Figure 21. AAIC reflectance image is on the left and the Material Id layer is on the right.
Images with a high percentage of clouds and cloud shadows will not perform well
with AAIC and the user should not expect removal of clouds and thick haze
that exists in the input image. Also, the current version of AAIC may not
generate accurately calibrated images in areas of high snow/ice cover.
The following figures show calibration results for two other images along with
their associated Material Id layers.
Figure 22. AAIC reflectance image is on the left and the Material Id layer is on the right.
Figure 23. AAIC reflectance image is on the left and the Material Id layer is on the right.
The following figures show the results when a poor calibration is achieved. The
reflectance image and associated Material Id layer are shown. In Figure 24, the
water in the image is identified as senescent vegetation, which is a definitive
indicator that calibration performance is poor. The user should attempt to find
another image covering the area of interest and not use the calibrated image for
any further processing.
Figure 24. Poor calibration results, AAIC reflectance image is on the left and the Material
Id layer is on the right.
In Figure 25, the calibration is questionable since there is a large area in the
open water that is being identified as soil. In addition, the land area does not
show much detail, which is indicative of an underperforming calibration. If
possible, the user should attempt to find a better image since any information
derived from this image will be suspect.
Figure 25. Poor calibration results, AAIC reflectance image is on the left and the Material
Id layer is on the right.