Entry Name: "UKON-Schlegel-MC3"
VAST Challenge 2017
Mini-Challenge 3
Team Members:
Udo Schlegel, University of Konstanz, udo.3.schlegel@uni-konstanz.de
PRIMARY
Alexandra Diehl, University of Konstanz,
diehl@dbvis.inf.uni-konstanz.de
Student Team: YES
Tools Used:
Tableau
Python
D3
SIZE: Satellite image zooming and exploration (Udo Schlegel, University Konstanz)
Approximately how many hours were spent working on
this submission in total?
50
May we post your submission in the Visual Analytics
Benchmark Repository after VAST Challenge 2017 is complete? YES
Video
Provide a link to your video.
The video is a mp4 file that is part of the submission: Link.
Questions
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.
With the tool, the user is able to
insert the image of the Boonsong lake into the different satellite images.
Afterwards he can drag, drop and scale it to any point and any size possible.
With a bit of exploration and scaling he can find the lake located in the
south-west. However, the tool already suggests the location based on a previous
exploration.
The image used of the Boonsong lake
is cropped to a height of 231 pixel and fits the 3000ft as good as possible
now. It is scaled to 0.14 to fit the location on the satellite images. The
orientation of the lake is the same orientation as the satellite images. So,
the satellite images are also oriented North-South.
This results in 3000 / ( 231 * 0.14
) = 92.7 = 93. As the satellite images have a size of 651 pixel it covers 60389
feet or 11.4 miles in height. If we assume it’s also the case for the width, we
get 93^2 = 8649 feet per pixel.
This discrepancies let us to the conclusion that a good
approximation for the area would be 93^2 feet per pixel with an error of 0.5% approximately.
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.
There are some features identifiable
on the images like lakes and streets, but there are more than just these
features.
1. Farmland
Figure 1 The pictures on top show the
plant health images (B4, B3, B2) from September 2015, June 2016 and
September 2016. On the bottom are the corresponding true color images.
On the top, circled in red, there
are some changes other time. With the plant health images, the shape and the
development, it is possible that these objects are fields of corn or crop. They
have a rectangular shape and different colors in different seasons.
2. Houses
Figure 2 The images show the plant health
(B4, B3, B2) from August 2014, June 2015, September 2015,
June 2016 and September 2016.
Along the farmland, there are
objects, which differ from the farmland and have a similar color to the
streets. So, you can suspect that these objects are houses of the farmers.
3. Rivers
Figure 3 The images show the flood (B5, B4, B2) images for August 2014, June 2015, September 2015,
June 2016 and September 2016.
On the flood images, it is possible
to find rivers (around an island in the middle) either floating into or out of
one of the lakes. It is only possible to distinguish these rivers on the flood
images, because of the small resolution.
4. Mountains
Figure 4 The top images and the first on the
left of the second row shows the snow (B1, B5, B6) images of March 2014, December 2014, February
2015 and December 2016. The one in the middle of the second row is the true
color image of December 2014 and the last one is the plant health (B4, B3, B2) of August 2014.
With these images, it is possible to
identify mountains in the satellite images. In the snow images, it is possible
to spot, which areas have more snow then others and through this can be speculated
to be higher than the surroundings. In the plant health images, red shows
vegetation and chlorophyll, so it is also possible to say that these mountains
aren’t high enough to consist only of a stone surface.
5. Streets and hiking paths
Figure 5 The two images show the true color and the
plant health (B4, B3, B2) of June 2015.
On the different images, it is
possible to recognize streets and hiking paths throughout the whole map. The
transit street in the middle is the most recognizable of all, but there are
more than just this one. Right next to the transit street and beneath the
lakes, there are a lot of different smaller streets and hiking paths, which are
lighter in the images.
6. Lakes
Figure 6 The two images show the flood images (B5, B4, B2) from March 2014 and December 2014.
There are five lakes on the
satellite images, easily identifiable with the flood images.
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.
1. Mining location changes
Figure 7 The images show the plant health images
(B4, B3, B2) from August 2014, June 2015, September 2015,
June 2016 and September 2016.
On the plant health images, there
are some locations, which change their color during the time. The color, the
two locations change from and to, is quite greenish and indicates, with the
help of the primer, vegetation color and mineral deposits. You could suspect
that there is a changing in locations for mining, which could interfere the
bird population.
2. Road expansion construction site
Figure 8 The images show the plant health images
(B4, B3, B2) from August 2014, June 2015, September 2015,
June 2016 and September 2016.
By examining the plant health images,
it is possible to identify that an already existing road in the beginning gets
expanded. It is quite small in early 2014 and gets larger in 2016.
3. Strange new mineral deposits
Figure 9 The images show the true color image
from February 2015, the plant health (B4, B3, B2) from June 2015, September 2015, June 2016,
September 2016 and the true color image from December 2016.
In the plant health images, there is
a strange new greenish field in the last two images of 2016. With the help of
the spectrum primer from the data, the greenish color means there is an either
a new vegetation or a new mineral deposit. This could also interfere with the
birds. It could be a plant, which is foreign and not liked by them or an
illegal digging location on a place, where they breed.
4. One of the lakes changes from
2014 to 2016
Figure 10 The images show the true color
images (B1, B5, B6) from December 2014 and December 2016 and the
corresponding snow images from December 2014 and December 2016.
One of the lakes changes from Winter
2014 to Winter 2016. In the true color images on top, it is already possible to
see that there is something strange in there. On the snow images, it is even
easier to spot differences. One of the lake has changed in general. On the left
side are the images from December 2014 and on the right, are the ones from
December 2016. The highlighted lake should have the same color as the one above
it. But a good part of it is more blueish. By the bands (B1, B5, B6) this means there is no ice there and the soil
mineral content changed. This could be the case because the lake was used to
deposit trash or other substances. Or it could just be a landslide. This could be another cause for the leaving of
the birds.
5. Global warming
Figure 11 The images show the snow images (B1, B5, B6) from December 2014, December 2015 and
December 2016.
During the winter or snow images, it
is possible to see a decreasing on the intensity of the reddish coverage. In
December 2014, red is a major color in the image. While in December 2016, there
is not much left of it. Red corresponds to general visible brightness and
should have a high value if there is white or more general snow. Even in the
image from November 2015, where a lot of clouds hide most of the surface, it is
more red.