I am a Research Assistant Professor in the Department of Natural Resources and the Environment at the University of Connecticut, Storrs, CT. My major research is to use remote sensing technology to understand global changes. Know more about me.
Cloud and cloud shadow detection in Landsat and Sentinel-2 data
Clouds and cloud shadows are the major noises for any applications of optical remote sensing data such as Landsat and Sentinel-2, especially in the ongoing era of dense time series data. Thus, we developed fully automated algorithms to detect clouds and cloud shadows, such as Fmask 4.0 and Cmask.
Consistency of Landsat time series
We examined several processing streamlines for improving Landsat data consistency, such as the impacts of data resampling, cloud/cloud shadow detection, BRDF correction, and topographic correction. Besides, Landsat 7's orbit has been drifting to an earlier acquisition time since 2017, which also affected the consistency. Learn more from this link and this link.
Temporal Landsat image compositing
We proposed a new Landsat image compositing algorithm (MAX-RNB) based on the maximum ratio of Near Infrared (NIR) to Blue band (RNB), and evaluated it together with nine other compositing algorithms. Learn more from this link.
Land disturbance product over CONUS
CONUS-wide land disturbance products were created based on the Landsat time series (this link).