I design and develop algorithms to fuse remotely-sensed satellite data from multiple imaging modalities such as UV, multispectral and hyperspectral VNIR, SWIR, high-res. RGB (including video) and develop standardized pipelines to generate different data products according to client needs.

Developed, expanded, and published a super resolution algorithm to spatially enhance low-resolution hyperspectral imagery using high-resolution multispectral imagery. Used CUDA libraries to train deep learning-based algorithms on the GPU using airborne data: hyperspectral, multispectral, and RGB. Performed atmospheric compensation/corrections on radiance data using MODTRAN to calculate reflectance images and spectral maps of various scenery, including plumes, vegetation, mineral-rich scenes (e.g., Cuprite). Modeled and simulated noise in camera systems, sensors, and images. Performed data cleaning on hyperspectral and multispectral datasets including de-noising, orthorectification, image registration, bandwidth selection, image stitching, histogram equalization. Utilized hyperspectral images for tasks such as: classification, target detection, object recognition, image segmentation, spectral unmixing, photogrammetry. Simulated high- and low-resolution images for image fusion as part of Wald’s Protocol in algorithm comparisons. Published a paper in the Journal of Applied Remote Sensing, April 2023 and presented at SPIE-Defense and Commercial Sensing (2023, 2022, 2021, 2020), IEEE-WHISPERS (October 2022).