SHARP logo Hope College, Holland, Michigan
All About SHARP applicants

projects

program directors

 

 

All About Sharp
Project details

Dune Imaging Research

Project Full Title:

Remote Sensing of Lake Michigan Sand Dunes to Study Dune Evolution and Vegetation Dynamics

Project Mentor(s):

Pearson,Paul

Project Mentor(s) EMail:

pearsonp@hope.edu

Project Start Date:

5/31/2019

Project End Date:

7/25/2019

Project Description:

This interdisciplinary project will incorporate the disciplines of applied mathematics and geology. Students applying to this project will be doing research in applied mathematics with Professor Paul Pearson. Sand dunes are dynamic environments that change rapidly as wind and vegetation act to erode and stabilize them. This project will study the earliest stages of open dune evolution by studying the dynamics of the vegetation community. In the Saugatuck Harbor Natural Area (SHNA) dune system, we will track small-scale changes in vegetation over time. We will collect ultra-high spatial resolution aerial photos from a small unmanned aircraft vehicle (a.k.a. drone), calibrate it with a small number of field measurements of above-ground biomass, and automate the construction of accurate above-ground biomass maps for all of SHNA from this data using deep convolutional neural networks from machine learning. Students will learn the theory of artificial neural networks (applied math), how to program them in Python (computer science), image processing (computer vision), and field work (geology). Students are not expected to know any of these things before the summer. Students will need to know linear algebra (through Math 231) before the research starts. Knowing partial derivatives and gradient vectors is a bonus, but not necessary. Professor Paul Pearson plans to apply for Nyenhuis and SFF grants to support students on this project. These grant applications are due in early January and require students to write a brief one-page statement of interest as part of the grant application. Consequently, students who are interested in this project should talk to Professor Paul Pearson in person in early December 2016. Contingencies: Professor Paul Pearson has already applied for a Michigan Space Grant Consortium (MSGC) Research SEED grant to pay for aerial photos by drone and for a new computer with fast GPUs specifically designed for deep machine learning. If the project does not receive funding from MSGC for aerial photography, it may not run (or it may get transformed into a project that gathers and analyzes a different kind of data). Professor Ed Hansen (geology) and a student are planning to help with the field work, but Professor Ed Hansen may only be available for a limited number of days during the summer (he may unavailable if he does research in Sweden). Professor Paul Pearson is somewhat flexible about the beginning and ending dates of the project.

External Link:

BACK

 

 

© 2007-2010 Hope College, Holland, Michigan, USA. All rights reserved.