Temporal Hyperspectral Representation Learning for Biotechnology
Self-supervised temporal hyperspectral representation learning for biotechnology.
projects
Research and software, from learning representations of Earth to simulating the sensors that observe it.
Scene Constructor: Large-Scale DIRSIG Scene Simulation
Automated scene construction for Landsat-scale simulations and sensor trade studies.
Landsat Split-Window Surface Temperature Uncertainty Framework
Estimating atmospheric water vapor and propagating uncertainty in Landsat split-window surface-temperature products.
3D Point Cloud Voxel Content Estimation in Simulated Forests
Estimating mixed material percentages inside forest voxels from 3D position and LiDAR intensity.
SPLASH: Multimodal Drone Hyperspectral Fusion
Aligning VNIR and SWIR pushbroom imagery from separate drones through global, row-wise, and elastic correction.
Snap Bean Yield Modeling and Harvest Scheduling
Connecting greenhouse spectroscopy with drone-based field observations for yield prediction and crop maturity assessment.
Jostar: Open-Source Feature Selection Library
Nine feature-selection algorithms for regression and classification, with a familiar Scikit-Learn-style interface.