Entry Name:  "UIC-Marai-MC3"

VAST Challenge 2017
Mini-Challenge 3

 

 

Team Members:

Bartosz Kupiec, University Of Illinois at Chicago, bkupie2@uic.edu     PRIMARY
Vijayraj Mahida, University Of Illinois at Chicago, vmahida2@uic.edu      
Timothy Luciani , University Of Illinois at Chicago, tlucia2@uic.edu      
Andrew Burks , University Of Illinois at Chicago, aburks3@uic.edu      
G.E. Marai , University Of Illinois at Chicago, gmarai@uic.edu

Student Team:  YES

 

Tools Used:

GIMP 2

JavaScript

d3.js , resemble.js, bootstrap.js

 

Github link

VAST Challenge 2017 - Mini Challenge 3 was developed by undergraduate researchers (REUs) at the Electronic Visualization Lab, University of Illinois at Chicago

 

Approximately how many hours were spent working on this submission in total?

100+ hours

 

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2017 is complete? YES

 

Video

https://www.youtube.com/watch?v=hRmjpg-hPzI&feature=youtu.be

 

 

 

Question 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.

 

The CSV files were converted to their respective RGB images, which were then used to find the Boonsong Lake’s position, given the picture with information about the lake (whose orientation was north-south and 3,000 ft. long). The clearest image (i.e. the one with minimal cloud cover and visible roads) seemed to be from data ‘image02_2014_08_24’. The satellite image was scaled to 7812 x 7812 pixels, then we took the original lake image and overlaid it on top of the satellite image. Upscaling was done on the image, so to preserve the lake size information. The lake image’s length is 347 pixels, which is about how much the lake takes up (where 347 pixels is 3,000 ft). Thus, one pixel spans 8.6455 feet, which means one pixel covers 74.745 ft^2. Overlaying the original image on top of the satellite images we were able to get, with a high degree of certainty, the scale (1 px = 74.745 ft^2) and orientation of satellite images (north-south).

 

lakeActuallakeSize

lakeLocation

Question 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 have created a web-based image analysis tool that combines small multiple views of satellite images, linked semantic zooming and image intensity histograms, along with filter controls, that allows us to visualize changes in the Preserve area.

 

 

 

There are several features that we identified by analyzing these images. The main features are five lakes, farmland (possibly campgrounds), parking lots, and roads. Using further image processing, we were further able to trace roads, based on pixel value similarity across 12 rendered images.  

 

With the multiple views we can see changes in the Preserve over the same season between 2014-2016: plant health, cloud cover, and snow-ice become readily apparent.

 

The “zoomed view” tool under the “image comparison” feature of our software enabled us to see that on 9/6/2016 and 12/19/2016 there was snow cover present in the area where the two major roads intersect, but the overall shape of that area remains the same. We can observe this because during September 2016 and December 2016 the shape of the roads and lot are roughly the same. Some sort of structure must be there, most likely camping grounds or parking. There appears to be sensor errors on the right side of the images across all bands except on 9/6/2016. bands.

 

 

 

 

 

Notes:

The following are features that we can identify with a high level of certainty:

 

 

 

 

 

 

 

 

Question 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.

Using the small multiples view to see the changes of the preserve over time and combining the image comparison tool we are able to see clear plant health changes over the time from 06/24/2015 to 09/12/2015. The point of interest is Lake D, where we suspect that there might be a case of possible chemical pollution, or perhaps due to a campsite of some sort. Also, in the RGB images the shadows of the clouds can be observed.

In November (of year 2016 and 2015) the preserve is very cloudy which would make sense since it’s close to winter time.

In the flood-burn mode the differences in the area of interest (see second image) become apparent as well. The percent difference between images is displayed on the bottom of our interface, so we can clearly see that some kind change has occurred in that time frame outside of the regular cloud coverage difference.

 

Notes:

           06/24/2015 - 09/12/2015.