2017 Annual Indiana GIS Conference – GIS Student Poster Award Winners

May 10, 2017, Bloomington, IN – IGIC is pleased to announce our GIS Student poster award winners at our Annual Indiana GIS Conference.  The IGIC GIS Poster Competition provides an excellent forum for GIS students to present their work to the Indiana GIS community. All current undergraduates and graduate students from Indiana universities are eligible to join the poster competition. Students are invited to submit illustrations of how GIS is used.

Judging Criteria:

Introduction and Methodology
• Statement of objectives
• Relation to previous research
• Research design
Results
• Substance of findings
• Relation of findings to objectives
• Conclusions
Design
• Organization
• Visuals

2017 Student Poster Competition Winners:

First Place ($250): Yuanfan Zheng, PhD Student, Indiana State University, Terre Haute, IN
“A Hybrid Approach for Three-Dimensional Building Reconstruction in Indianapolis from LiDAR Data”

Second Place ($150): Samapriya Roy, PhD Student, Department of Geography, Indiana University, Bloomington, IN
“Deep Time Stack analysis of Coastal Land loss: Case Study of Mississippi Delta using Earth Engine”

Third Place ($100): Ben Scholer, Undergraduate, Computer Science, Purdue University, West Lafayette, IN
“Indiana Water Monitoring Inventory and Map”

Details about the Winning Posters:

First Place: Yuanfan Zheng, PhD Student, Indiana State University, Terre Haute, IN
A Hybrid Approach for Three-Dimensional Building Reconstruction in Indianapolis from LiDAR Data

3D building models with prototypical roofs are more valuable in many applications than 2D building footprints. Although using the high resolution LiDAR point clouds or InSAR data can improve the accuracy of 3D building reconstruction, it requires a large amount of data. In particular, the reconstruction of mixed and complicated roof structures within large-scale areas is computationally intensive and lacking in effective approaches. Existing studies of the reconstruction of 3D city models from LiDAR-derived nDSM or high-density point cloud data with or without additional data sources included three major categories of strategies: data-driven (bottom-up), model-driven (top-down), and hybrid approaches.
According to the existing studies mentioned above, it can be concluded that the hybrid approach can produce higher reconstruction quality than using data-driven and model-driven approaches separately, as it is more flexible than the model-driven approach and can produce much more complete results than the data-driven approach. However, even more complete planes can be generated by the hybrid approaches’ accuracy of plane boundaries, which still can be limited by the precision of algorithms used in data-driven approaches, which detect the edges and ridges.
The proposed hybrid approach for the LoD2 3D city model reconstruction from the LiDAR data was tested in the CBD of the City of Indianapolis, USA. A total of 519 buildings, including tall commercial buildings and residential buildings with different roof shapes, were reconstructed.

Second Place: Samapriya Roy, PhD Student, Department of Geography, Indiana University, Bloomington, IN
Deep Time Stack analysis of Coastal Land loss: Case Study of Mississippi Delta using Earth Engine

In the next few decades as urban migration and population pressure increases worldwide, the landscapes sustaining these increasing loads will change as well. Delta systems are unique in this domain because while they consist of only 3% of the world’s land area, they have one of the highest population densities manifold that of other landscapes. The Louisiana coast has been facing both natural and urban degradation and has been losing land rapidly with current estimates at about 5350 square meters of land every hour (approximately the same as a football field). This project looks at deep time series datasets from 1983-2016 to get better estimates of land loss quantity and pattern of loss.
The methodology builds upon a pipeline of image processing which is built on the distributed computing environment provided by Earth Engine. A cloud free composite is created using a percentile reducer method. The next step is to look at index approaches or multi-dimensional clustering approached (in this case Modified Normalized Difference Water Index and Constrained Spectral Unmixing). For both of these methods we get an index based gradient out and Spectral Unmixing further allows us to look at sub pixel classifications.
This project allows us to look at an unprecedented amount of data analyzing over 10,000 Landsat tiles and more than ~1.30 billion pixels. Distributed computing allows us to overcome issues with sorting and creating composites along with the application of pixel based algorithms over large areas. The results confirm that we have lost over 1200 square kilometers of land from 1986-2016 within just our area of interest.

 

Third Place: Ben Scholer, Undergraduate, Computer Science, Purdue University, West Lafayette, IN
Indiana Water Monitoring Inventory and Map

This poster describes a project that upgrades the old Indiana Water Monitoring Inventory (IWMI) site. The original inventory and map was built in conjunction with the Indiana Water Monitoring Council. Using this improved site, users are able to do similar functions, using a more streamlined, modern interface. The new map represents a continuation of the original software design and development project, to provide newer technology, backup and prototyping capability. The map is drawn using the Google Maps Web API V2, and to improve the performance, the design was switched to using Marker Clusters, which organize markers in groups, and drastically speeds up rendering.

Speak Your Mind

*

*