RESEARCH

Marine and Coastal Debris Detection and Flow Prediction

Dec. 2019 — Present

Detect and track marine plastic with multispectral images provided by several satellite platform including Planet labs, Sentinel, GOES images associating with local coastal data. Implemented algorithm with geospatial library and machine learning tools on python platform

Geographical Knowledge Learning with High Resolution Earth Imagery

Aug. 2019 — Present

Combining GPS and GIS information to object detection algorithm for high resolution satellite image data. Provides the detail spacial distribution and extract semantic features applying anomaly detection algorithm. which has its capability to save annotating costs for supervised learning.

Satellite’s On-board Image Feature Detection by Neural Network

Apr. 2016 — Mar. 2019

A sensor developed and launched in orbit mounted on JAXA satellite RAPIS-1. We demonstrated on-orbit image recognition using on-board computer, which aims to be applied to attitude sensors and real-time Earth monitoring