Researchers from the University of Lisbon, including Leonel Corado, Sérgio Godinho, Carlos Alberto Silva, Juan Guerra-Hernández, Francesco Valérioa, Teresa Gonçalves, and Pedro Salgueiro, have developed a new tool aimed at improving the accuracy of geolocation data from the Global Ecosystem Dynamics Investigation (GEDI) LiDAR mission. This tool, named GEDICorrect, is designed to enhance the reliability of GEDI data for applications such as biomass modeling, data fusion, and ecosystem monitoring.
GEDI, a NASA mission launched in 2018, uses LiDAR technology to measure the height and structure of Earth’s forests and other ecosystems. Accurate geolocation is crucial for the effective use of this data, as residual errors can significantly impact the accuracy of derived metrics. The researchers developed GEDICorrect to address these issues, providing a flexible and computationally efficient framework for correcting geolocation errors at various levels: orbit, beam, and footprint.
The GEDICorrect framework integrates existing GEDI Simulator modules and extends their functionality with flexible correction logic, multiple similarity metrics, adaptive footprint clustering, and optimized input/output handling. By using the Kullback-Leibler divergence as the waveform similarity metric, the tool improved canopy height accuracy from an R-squared value of 0.61 to 0.74 with orbit-level correction, and up to 0.78 with footprint-level correction. This reduction in error translates to a decrease in the root mean square error (RMSE) from 2.62 meters to 2.12 meters at the orbit level, and 2.01 meters at the footprint level. Terrain elevation accuracy also saw improvements, with a reduction in RMSE by 0.34 meters relative to uncorrected data and by 0.37 meters compared to the GEDI Simulator baseline.
In terms of computational efficiency, GEDICorrect achieved a significant speedup over the GEDI Simulator. In single-process mode, the runtime was reduced from approximately 84 hours to 35 hours, and with 24 cores, the task was completed in about 4.3 hours, representing an overall improvement of roughly 19.5 times. This efficiency makes GEDICorrect a practical tool for large-scale data processing tasks.
The practical applications of GEDICorrect for the energy sector are notable. Accurate geolocation data is essential for renewable energy projects, particularly in the planning and monitoring of wind and solar farms. For instance, precise elevation and canopy height data can aid in site selection and assessment of potential energy yields. Additionally, improved data accuracy can enhance the integration of LiDAR data with other geospatial datasets, supporting better decision-making in energy infrastructure development and environmental impact assessments.
The research was published in the journal Remote Sensing of Environment, providing a robust and scalable solution for improving GEDI geolocation accuracy while maintaining full compatibility with standard GEDI data products. This tool is expected to enhance the reliability and utility of GEDI data for various applications in the energy sector and beyond.
This article is based on research available at arXiv.