You are here

Geodetic research team excels in international competition

Researchers from Ohio State's Geospatial Data Analytics Group were multiple track winners of the 2019 Institute of Electrical and Electronics Engineers (IEEE) Data Fusion Contest. The IEEE states that this annual, worldwide contest "aims to promote research in semantic 3D reconstruction and stereo algorithms using machine intelligence and deep learning applied to satellite images." Teams from around the globe submitted 1,100 entries for the two phase contest that began in January of 2019 and concluded on March 22, 2019.

Rongjun Qin, PhDOhio State's team, led by Professor Rongjun Qin, consisted of postdoctoral researchers Xu Huang and Changlin Xiao as well as visiting PhD student Wei Lu. Competing against many larger teams, Qin and company won two of four individual competitive tracks in the contest. The Pairwise Semantic Stereo Challenge required participants to predict semantic labels and stereo disparities in images and data provided by the organziers while the Multiview Semantic Stereo Challenge asked teams to predict semantic labels and a digital surface model (DSM) based on provided data samples.


Professor Qin was quite pleased with the team's finish as no institute from the United States had placed in any of the IEEE contests held during the last five years. "This puts Ohio State at the forefront of machine learning and 3D reconstruction competence for satellite images" he said. "Historically, the United States is not as strong in this arena as the usual winners from Europe and China." In July of this year, Qin's team will present their winning methods to the worldwide community at the 2019 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) in Yokohama, Japan.

(L to R) Geospatial Data Analytics Group members Xu Huang, Changlin Xiao and Wei Lu

Qin's research, which incorporates remote sensing, photogrammetry and computer vision, has many practical applications in civil and environmental engineering, including 3D city and landscape modeling and environmental monitoring and disaster responses.