Saurabh Prasad, assistant professor of electrical and computer engineering at the UH Cullen College of Engineering, has won NASA’s New Investigator Award in Earth Science. Out of over 130 proposals submitted to the New Investigator Program (NIP), only 21 were selected – two of which were from the University of Houston Cullen College of Engineering (click here to read about civil & environmental professor Hyongki Lee’s NIP award!). The New Investigator Program was established by NASA in 1996 to support outstanding scientific research and career development of scientists and engineers at the early stages of their professional careers.
Prasad is a leading expert on signal processing and image analysis who joined the Cullen College faculty in 2012. He leads a research group on geospatial image analysis in the Cullen College’s electrical and computer engineering department. He is also affiliated with UH’s National Center for Airborne Laser Mapping, or NCALM. As a researcher, Prasad develops cutting-edge image analysis algorithms for remotely sensed geospatial data collected from multiple sensing platforms, including airborne (sensors aboard an aircraft) and spaceborne (sensors aboard satellites in orbit), ground based imagers, and much more.
This latest award focuses on developing fundamentally new approaches (algorithms) for advancing state-of-the-art geospatial image analysis. Prasad said his hope is that the basic image processing research he is conducting through the NASA NIP award will substantially enhance the capability of scientists to better interpret data from a diverse array of geospatial imaging and sensing modalities, such as hyperspectral imaging (imaging spectroscopy), light detection and ranging (LiDAR) measurements and synthetic aperture radar (SAR) measurements. In this project, however, the algorithms he develops will largely be validated by hyperspectral imagery and other passive optical imagery (such as images taken by orbiting satellites) acquired by NASA.
This research will be built upon three key aspects. The first goal of the research is to develop advanced algorithms and disseminate them to the broader research community. The second goal is to develop novel protocols for rapid ground-reference data collection using hyperspectral imagers to support remote sensing analysis needs. The third and final goal is to validate these approaches by employing them to study the ecology of the coastal wetlands in the Gulf of Mexico over wide spatial scales.
Currently, our ability to understand the complex ecosystem of the gulf coast wetlands relies on teams of scientists who must travel by boat, from point to point, to gather data on the local plants, animals and hydrology (the movement and quality of water) in the wetlands. This information, as difficult and tedious as it is to collect, is vital to our understanding of climate change and its overall impact on the environment.
But now, thanks to Prasad’s effort, scientists would be able to employ satellite and aerial imaging data in conjunction with field measurements for a much more robust understanding of coastal wetlands at various levels of detail – from very high resolution ground-based hyperspectral imagery that quantifies local processes, to wide-scale aerial and satellite imagery that can inform scientists on holistic trends related to ecosystem health.
The algorithms Prasad is currently developing can take enormous data sets from geospatial sensors and turn them into maps that accurately characterize the ground cover. Prasad explained that his mapping algorithms could be utilized to quantify metrics such as vegetation health, water quality, changes in vegetation cover, and sediment deposits over time, among other indicators of ecosystem health.
“If you’re able to visualize and understand what’s going on at a wide geographic scale using remote sensing reliably, then such data and methods can be indispensable tools to track ecosystems over time,” explained Prasad.
One challenge that scientists often face with remote imaging and sensing technologies is that these technologies have evolved faster than the capacity of traditional off-the-shelf software to process and analyze such data effectively. In particular, remote sensing can now provide users with extremely robust and complex data sets – but to effectively analyze such high-dimensional data sets is still a big hurdle, and there is a need for algorithms to address this challenge.
Luckily, the software suite that Prasad and his team develop to analyze the data from remote sensing technologies will be made available to the scientific community. Prasad said he hopes that making this software available to the broader research community will enable “rapid, robust understanding of ecosystems (such as coastal wetlands) over wide geographic scales.”
The algorithms that will be developed in this project will be scalable and transferrable to a diverse variety of geospatial data. Additionally, Prasad will also continue on his existing efforts to develop data fusion algorithms to “fuse” information from multiple imaging modalities. His previous (ongoing) NASA project on data fusion (specifically, fusion of hyperspectral imagery with LiDAR data) bears synergy with the new research in the NASA NIP project. By combining multiple data sets, Prasad is able to conduct a much richer image analysis.
Moreover, the software Prasad develops could have applications that go far beyond the mapping of coastal wetlands. “These algorithms could be used for many different types of applications, including biomedical image analysis,” said Prasad.