TCarta Leverages AI Technology and ICESat-2 Data to Create a G-SDB Product

TCarta Marine, a global provider of hydrospatial solutions, has introduced a Global Satellite Derived Bathymetry (G-SDB) product line developed with a new seafloor depth measurement technique that leverages Machine Learning and NASA ICESat-2 laser data.

The first G-SDB offering covers the entire Red Sea – available now – with additional data sets rolled out through the end of this year.  

The commercial TCarta G-SDB data sets and the state-of-the-art seafloor measurement workflow that produces them were made possible through a Small Business Innovation Research Grant from the National Science Foundation (NSF).   

“The new satellite-derived bathymetry technology extracts seafloor measurements by integrating multiple SDB algorithms and sensor types at scale and over broad geographic areas with a degree of confidence in data accuracy not previously possible,” said TCarta president Kyle Goodrich.   

G-SDB data sets contain bathymetric measurements to depths of more than 30 meters, depending on water clarity, at 10-meter resolution. The depth values for every 10m pixel are the combined result of numerous measurements, resulting in accuracy within 10% of depth or less, and providing a seamless water bottom surface map. G-SDB will be available globally for all oceans and seas, as well as large freshwater lakes where water conditions permit.   

“Continually updated high-quality bathymetric products available for immediate purchase will be game-changers for engineering companies planning near-shore infrastructure projects for oil & gas, port development, pipeline construction, and many others,” said Goodrich. “TCarta’s traditional customers in GEOINT, environmental management, and coastal zone monitoring will value the G-SDB products as well.”   

TCarta launched Project Trident with NSF funding in 2018 with the goal of refining traditional satellite-derived bathymetry technology to extend its application into areas where it had not typically been successful, usually due to the turbidity or clarity of the water column. TCarta developed the new method using machine learning to iteratively evaluate Sentinel 2A/B multispectral satellite images, and even individual pixels within images, to select the sharpest and clearest ones for application of SDB extraction.   

“Thanks to the power and speed of cloud computing, we run the extraction algorithm repeatedly and on multiple satellite images acquired over the same geographic area on different dates,” said Goodrich. “This dramatically increased the accuracy confidence in each depth measurement and minimizes data gaps.”   

To further enhance the accuracy of the SDB measurements, TCarta has developed a first-of-its-kind AI-based technique for leveraging ICESat-2 data to train the SDB algorithm and validate results. Designed for polar ice elevation and tree canopy measurements, the ICESat-2 satellite carries a laser that captures remarkably accurate bathymetric data. The TCarta technology is believed to be the first commercial bathymetric application of ICESat-2 data.  

Cloud computing also makes the entire G-SDB process faster. In just one week, TCarta evaluated and processed more than 17,000 multispectral images acquired since 2015 by the European Sentinel 2A/B satellites to create the Red Sea G-SDB data set. At current production rates, TCarta expects to have the worldwide G-SDB data set completed by late 2021.