We are attempting to obtain the best stream outline of a portion of Duck Creek that we can in an affordable and scientifically supportable way. As our first week of research is winding down, we have been able to draw some conclusions on potential software and techniques that could be used to assess the stream outline of a region of Duck Creek. We began the week with some quick introductions then dove right into brainstorming. After brainstorming we took a field trip to Duck Creek and quickly realized that some of our ideas could already be eliminated due to the width of the stream.
One of the more popular techniques we have been researching is photogrammetry—the use of photographs in mapping. To explore this idea further, we constructed a fake “stream” (see picture) on the tile floor of our lab with tape. Pictures were taken of the stream at many different locations and angles. We used these pictures to explore a wide variety of photo processing software (in some cases spending hours trying to understand and utilize a single software) in order to determine the best way to correct for perspective and distortion caused by the camera. This is necessary to determine accurate measurements of the stream outline.
In particular, a photo processing software called Hugin was used in an attempt to correct for perspective and stitch images together. Kyle was able to stitch images together (see picture) but each of us, including the professors, fought with this software for hours without having much luck in terms of perspective correcting. However, Dr. Forrest Stonedahl was able to prove that Hugin could potentially be very useful after being the only one to effectively accomplish this. We are excited to further explore this software in the stream setting and have been brainstorming ideas that would make Hugin even more successful for our project in particular.
A computer program called Matlab was also explored for an afternoon to create various types of 3D images. This technique could further be explored to create a 3D image of the Duck Creek stream outline once sufficient data is gathered. We have also been searching literature for relevant articles and read a paper by Mandelbrot, which discussed how the length of the coast of Britain was fractal in nature and would be measured to be different lengths depending on the measurement resolution used. This is applicable to stream outlines, which may not be fractal, but would be measured to be different lengths depending on how much detailed is included in the map.
The rest of the week will be spent testing more software and searching for alternate methods in research articles. We look forward to taking the cameras out to the Duck Creek Stream and possibly combining photogrammetry with grid making to begin obtaining data next week. We are keeping our own blog, which you could visit if you would like more pictures, information, or general stream of consciousness from our group than is included above: http://streamoutlines.blogspot.com/ .