An often-cited example of how public spending for scientific research can yield commercial dividends is the NASA moon landing mission. The forerunners of a number of products that today are ubiquitous, such as the personal computer, first took flight aboard Apollo. What might not be as well-known is the fact that NASA continues to develop technologies that can offer significant commercial potential. According to Dr. Scott Hubbard, Professor of Aeronautics and Astronautics at Stanford University, every $1 invested in NASA produces around $7 to $8 in goods and services for the economy. And these technologies often find commercial applications far removed from the missions for which they were originally developed.
NASA’s Goddard Space Flight Center (GSFC) has developed numerous technologies to support space science missions such as the Hubble Space and James Webb Space Telescope. These telescopes rely on a technique called image processing — very basically, the use of software to obtain the most information possible from a raw image. In addition to its obvious value in space exploration and astronomy, image processing is also critical to Earth science involving the monitoring of our planet’s surface features, in fields such as agriculture and energy production. And many of these image processing technologies can be applied to terrestrial applications by licensing them through GSFC’s Innovative Partnerships Program (IPP).
Remote Sensing Applications
Remote sensing is one area in which image processing plays a critical role. As the name implies, remote sensing involves obtaining information about an object without coming in physical contact with it. This data acquisition typically entails the use of the electromagnetic (EM) radiation, both within and outside optical wavelengths.
Remote sensing uses a variety of image capture techniques (often from aerial platforms such as satellites and aircraft), including optical, hyperspectral, multispectral, radar (including synthetic aperture radar, or SAR), LIDAR, and others. For instance, a technique increasingly employed in agricultural and forestry applications is hyperspectral imaging. This involves capturing a broad portion of the electromagnetic spectrum, and then processing this data to produce a visible image. The results can then be enhanced to amplify features of interest, such as pest infestations, weeds, disease, climatic conditions, vegetation health, and many other variables important to agricultural and forestry applications. Hyperspectral imaging thus combines spectroscopy and imaging technologies, and takes advantage of the fact that every chemical has its own unique spectrum. In addition to detecting agricultural problems, hyperspectral imaging can help predict crop yields (thereby helping farmers with their economic planning). Hyperspectral imaging can also be used in so-called “precision farming,” to help guide autonomous farm machinery to address areas in which their services are needed. The latter technique can also be applied to precision forestry.
Remote sensing technologies can also be applied to fields as diverse as the study of atmosphere (water vapor, cloud properties, aerosols), ecology (chlorophyll, leaf water, cellulose, pigments, lignin), geology (mineral and soil types), coastal waters (chlorophyll, phytoplankton, dissolved organic materials, suspended sediments), snow/ice (snow cover fraction, grain size, melting), biomass burning (sub-pixel temperatures, smoke) and commercial mineral exploration. For example, SAR is frequently used for petroleum exploration. SAR employs reflected radar pulses which are processed to determine the composition of the object that caused the reflection. Mining is also an important market for remote sensing, especially for multispectral, SAR, and hyperspectral based images, data, and value-added-services.
Hierarchical Image Segmentation (HSEG)
Of course, these and other remote sensing technologies are highly dependent on image processing. Capturing the raw EM data is only the first step; converting that data into a meaningful image can be a major challenge. This is one area in which GSFC has a great deal of applicable experience.
For example, GSFC scientists have developed a technology called hierarchical image segmentation (HSEG). This uses an algorithm that closely intertwines image segmentation via region growing, which finds spatially connected region objects with region object classification, which in turn groups sets of region objects together into region classes. This produces a segmentation hierarchy, or a set of several image segmentations of the same image at different levels of detail. Segmentations at coarser levels of detail can be produced from simple merges of regions at finer levels of detail. This allows an object of interest to be represented by multiple segments in finer levels of detail, merged into a surrounding region at coarser levels of detail.
HSEG was originally developed to enhance and analyze images such as those taken of Earth from space by NASA’s Landsat and Terra missions. Potential HSEG applications include topographical analysis, such as ice and snow mapping. Another application is “space archeology,” where satellite images are examined for subtle signs of previous human activity.
As a demonstration of how NASA technologies can be applied to markets far removed from their original purpose, HSEG has now been adapted to medical imaging. Bartron Medical Imaging has licensed HSEG to enhance its medical imagery products, allowing for quicker and more accurate identification of problematic tissues such as cancer. The technology has now been developed as MED-SEG, a tool to help specialists interpret medical images. In August of 2010, MED-SEG received clearance from the U.S. Food and Drug Administration (FDA) to be sold on the commercial market.
Recently HSEG has also been adapted for hyperspectral imagery. Another possibility is the fusing of spot LIDAR data with continuous coverage image data, which appears to be another application for which HSEG is suited. This continued development of HSEG offers some interesting licensing possibilities, for markets as diverse as mineral and petroleum discovery, forest management, and crop monitoring.
Multi-Wavelength, Multi-Beam and Polarization Sensitive Laser Transmitter
There is a lot of information to be found in even a single photon of light. To take advantage of this fact, GSFC has developed laser-based technology that can not only tell where an object is, but also help identify its physical characteristics as well. This technology may someday be used in applications as diverse as monitoring the movement and thickness of sea ice — or diagnosing tissues within the human body.
GSFC has been developing space-borne laser altimeters for a number of years, for missions such as ICESat (Ice, Cloud, and land Elevation Satellite). Launched in late 2003, ICESat helped scientists monitor ice sheet mass balance and other polar-specific environmental phenomena. ICESat’s successor, ICESat-2, is scheduled for launch in 2016. ICEStat-2’s objectives include quantifying polar ice-sheet contributions to current and recent sea-level change, quantifying regional signatures of ice-sheet changes, estimating sea-ice thickness, measuring vegetation canopy height, and monitoring tectonic plate movement
The receiver for this system can process the return beam with literally single-photon precision. This allows for unprecedented mapping accuracy. It also allows the system to obtain very specific data about the area being observed. Thus this technology could be applied to commercial applications such as precision mapping and remote sensing.
Of course, the NASA technologies discussed in this article represent only two examples. There are numerous others available for licensing, technologies that with a little creativity could be leveraged into new uses in environmental and energy production applications — and potentially a broad spectrum of other markets as well. To learn more about NASA image processing technologies available for licensing, contact the Goddard Space Flight Center Innovative Partnerships Program Office (see http://ipp.gsfc.nasa.gov/).