Entry Name: "VRVis-Matkovic-MC3"
VAST
Challenge 2017
Mini-Challenge
3
Team Members:
Michael
Beham, VRVis Research
Center, Vienna, Austria, Beham@VRVis.at
PRIMARY
Rainer
Splechtna, VRVis Research Center,
Vienna, Austria, Splechtna@VRVis.at
Silvana
Podaras, VRVis Research
Center, Vienna, Austria, podaras@vrvis.at
Denis
Gracanin, Virginia Tech, Blacksburg, VA, USA, gracanin@vt.edu
Kresimir Matkovic,
VRVis Research Center, Vienna, Austria, Matkovic@VRVis.at
Student Team: NO
Tools Used:
Matlab scripts: data processing and determining
differences between images.
Octave
scripts: data processing and images processing, e.g., edge detection.
GIMP:
image processing.
Google
calculator: numerical computations.
Approximately how
many hours were spent working on this submission in total?
80h
May we post your
submission in the Visual Analytics Benchmark Repository after VAST Challenge
2017 is complete? YES
Video
Questions
MC3.1 – Boonsong Lake
resides within the preserve and has a length of about 3000 feet (see the Boonsong Lake image file).
The image of Boonsong Lake is oriented
north-south and is an RGB image (not six channels as in the supplied satellite
data). Using the Boonsong
Lake image as your guide, analyze and report on the scale and orientation of
the supplied six-channel satellite images.
How much area is covered by a pixel in these images? Please limit your answer to 3 images and 500
words.
We identify
SIFT Features in RGB sensor data using Grayimage,
pattern in RGB GrayImage and RANSAC Algorithm (Figure 1). Guide image has the same
orientation like the pattern.
Figure 1: Feature matching. Image
registration is done using SIFT features + RANSAC. That results in a
transformation matrix. We use different scale parameter for SIFT features to
detect enough features.
We use the shape
of the lake to refine the transformation matrix (Figure 2).
Figure 2: Refining the measurement.
We measure
approximately 30 pixels for the lake which is equivalent to 3000 feet.
Therefore, the resolution is 3000 feet / 30 pixels which is equivalent to 100
feet/pixel (30.48 meters/pixel).
The size of
the satellite images is 651 pixels (both height and width) * 100 feet/pixel
which is equivalent to 65100 feet or 12.33 miles (both height and width).
MC3.2 –Identify features
you can discern in the Preserve area as captured in the imagery. Focus on image
features that you are reasonably confident that you can identify (e.g., a town
full of houses may be identified with a high confidence level). Please limit
your answer to 6 images and 500 words.
We start by using
RGB images (Figure 3), false-colored images (Figure 4) and vegetation indices (Figure 5) and then apply ratio
transformation (Figure 6).
Figure 3: RGB images.
Figure 4: False-colored images.
Figure 5: Vegetation indices.
Figure 6: Ratio transformations.
The
identified features are shown in Figure 7. We identified it using
false-colored images and measurements like NDSI. Figure 8 shows a questionable feature that
indicates freezing areas on a lake surface (freeze hole).
Figure 7: Identified features.
Figure 8: Freeze hole.
MC3.3 – There are most
likely many features in the images that you cannot identify without additional
information about the geography, human activity, and so on. Mitch is interested in changes that are
occurring that may provide him with clues to the problems with the Pipit
bird. Identify features that change over
time in these images, using all channels of the images. Changes may be obvious or subtle, but try not
to be distracted by easily explained phenomena like cloud cover. Please limit your answer to 6 images and 750
words.
Some easily detected phenomena include cloud cover
(and cloud shadow), seasonal changes (e.g. snow, changes in vegetation), and
sensor artefacts.
Preprocessing requires generating cloud and shadow maps (Figure 9).
Figure 9: Cloud and shadow maps.
We compare satellite images captured at similar time of
the year. Comparison by frame differences is done by calculating differences
for visible and invisible band combinations and then setting the differences,
resulting by clouds and images, to 0. Focus is on the interesting parts of the
image (Figure 10).
Figure 10: Frame differencing - creating the
difference image.
The resulting changes between 24 August 2014 and 6 September
2016 are shown in Figure 11.
Figure 11: Changes between 24 August 2014 and
6 September 2016.
Seasonal
vegetation changes between 26 June 2016 and 6 September 2016 are shown in Figure 12.
Figure 12: Seasonal vegetation changes between
26 June 2016 and 6 September 2016.