Advanced Computer Vision, Robotics, and Visualization Algorithms for Improving Planetary Exploration and Understanding University of Nevada at Las Vegas/NASA Ames Research Center, Space Technology Mission Directorate
Researchers from the University of Nevada at Reno (UNR), University of Nevada at Las Vegas (UNLV), and Desert Research Institute (DRI) have been working together on advancing NASA’s technologies for planetary exploration and understanding. The UNR team has developed a new technique for horizon line detection using machine learning to aid in determining rover location and orientation. They have also been working on detecting craters from orbital images by developing a novel methodology that employs convex grouping for extracting candidate crater regions and machine learning for verifying them. The UNLV team has developed new techniques using GPU boards as well as an interactive rock segmentation and quantification (RSQ) tool to reduce image noise. The RSQ tool allows users to select an area of interest, segment rocks of interest, and calculate various rock properties based on color, texture, and shape. The DRI team has employed image and elevation data analysis techniques for mapping
rover mobility over a landscape that exhibits variable surface lithology and geometries. We have strengthened our collaboration with NASA as a result of this NASA EPSCoR project and are now working with the Intelligent Robotics Group (IRG) at NASA Ames. We have been holding annual meetings with IRG to review our progress. One result has been joint publications in peer-review journals and conference presentations. A special session on “3D Mapping, Modeling and Surface Reconstruction” was organized at ISVC’14. Research results have been incorporated in student classes and seminars have been provided by invited NASA researchers. Highly qualified students have been involved in this project as research assistants or summer interns. We have leveraged project results to obtain new funding; to deploy UAVs for improving situational awareness of first responders.
Sample results illustrating our horizon line detection approach: City data set (row1), Basalt Hills data set (row2) and Web data set (row 3 through 5). Detected horizon lines are highlighted in red/green.
University of Nevada, Reno
NASA EPSCoR Stimuli 2014 -15
Dr. George Bebis, Science PI, Nevada System of Higher Education
Terry Fong, NASA Technical Monitor, Kennedy Space Center
Published on Dec 14, 2015
NASA Office of Education’s Aerospace Research & Career Development (ARCD) is pleased to release NASA EPSCoR Stimuli, a collection of univers...