Estimating Canopy Height Change Using Machine Learning by Coupling WorldView-2 Stereo Imagery with Landsat 7
A study recently published in the International Journal of Remote Sensing titled Estimating canopy height change using machine learning by coupling WorldView-2 stereo imagery with Landsat-7 data discusses the challenges in mapping changes in forest canopy height over large areas using limited data from field measurements and Airborne Laser Scanning (ALS).
Co-authors of the study were Jian Wang, a Ph.D. student at The Ohio State University, along with Professors Desheng Liu and Steven M. Quiring from the Department of Geography and principal investigators at the Byrd Center and Associate Professor Rongjun Qin, who holds appointments at the Department of Civil, Environmental and Geodetic Engineering, Department of Electrical and Computer Engineering, Translational Data Analytics Institute and Environmental Science Graduate Program.
This research highlights the potential of very high-resolution (VHR) satellite stereo imagery to estimate digital surface models (DSMs) and explores its capability and potential to estimate canopy height models (CHMs) and canopy height change (CHC). The study evaluated stereo-based CHMs and CHC from WorldView-2, a commercial Earth observation satellite, over five woody parks in Columbus, Ohio, and integrated stereo-based CHM with vegetation indices from Landsat 7 to improve CHM and CHC estimation with machine learning methods. The results show the limitations of stereo imagery alone in estimating CHMs and CHCs but combining remote-sensing structural and spectral information can improve the estimations.
Learn more about this study by visiting the Research Article online.