Earth Imaging Journal

Home

About Us

Articles

Press Releases

Image Gallery

Market Guide 2010

Media Kit

Free Subscription

Industry Resources

Industry Directory

NASA Links

Missions/Launches

Calendar



	Homeland security involves a variety of activities that protect life, property and critical infrastructure. Remote sensing and related geospatial technologies play a vital role in these activities by detecting, identifying, mapping and monitoring the effects of natural and man-made hazards and disasters, as well as enhancing emergency management through time-sensitive damage assessments. Much of this discussion focuses on remote sensing for assessing vegetation damage and soil contamination, including artificial radionuclide contamination, along with damage to urban areas and other human settlements. Environmental Damage Assessment A key component of homeland security lies with securing and maintaining a healthy, sustainable environment to ensure sufficient agricultural production and food security. Remote sensing is valuable for agricultural monitoring and modeling, as well as for assessing climate change, which may significantly affect food and water supplies. For example the U.S. Department of Defense recently released a report examining the link between climate change and U.S. national security. The report noted the potential for climate change to destabilize the international geopolitical environment, possibly resulting in future conflicts and wars based on related food and water shortages. Regardless of the feasibility of such a scenario, remote sensing can help assess vegetation stress or damage stemming from a variety of sources. One agricultural threat to homeland security and international stability may take the form of Ug99, a virulent strain of black stem rust fungus (Puccinia graminis). Discovered in Uganda in 1999, the fungus has since spread via airborne spores, though it also can be transported by travelers. The Puccinia graminis fungus affects wheat, which is significant because wheat feeds more people than any other food source. Research into rust fungus-resistant varieties is being conducted, and remote sensing can be used to detect and monitor wheat rust. Another important agricultural threat is drought, which is among the costliest natural hazards in the United States. Unlike some other hazards, drought develops slowly; however, it can exert severe environmental, economic and social consequences, the effects of which can persist over a long time period. Other types of natural disasters also can adversely affect vegetation communities, including floods, hurricanes and severe windstorms.




	Many remote sensing studies have addressed vegetation abundance or health/stress mapping by using vegetation indices, such as the normalized difference vegetation index, to estimate relative abundance and photosynthetic activity of green vegetation. Hyperspectral imaging is particularly useful for such applications, where hundreds of contiguous bands are available for information extraction. Many of the concepts used to assess natural disaster-induced vegetation damage also apply to man-made scenarios. However, special consideration is given here to nuclear/radiological event-induced vegetation damage and soil contamination. If artificial radio-nuclides are released from a nuclear facility—either accidentally or deliberately as the result of a terrorist strike—gamma spectrometric sensors can be used to quickly and broadly map nuclide-specific contamination. In situ and airborne gamma spectrometric sensors can be used to survey point releases from nuclear facilities, monitor radioactive fallout over large regions and detect lost or stolen radioactive sources. Mobile or fixed gamma-ray sensors can provide field data. Mobile in situ gamma spectrometers can be operated on foot or by car to sample at specific locations as required. Fixed monitoring networks can continuously measure the gamma dose rate. One example of a sensor network is the SensorNet Program, which was designed and developed by Oak Ridge National Laboratory (ORNL) and partners. SensorNet began with the concept of mounting wireless sensors on static and mobile platforms for emergency response. SensorNet involves developing data-processing and information-delivery methods for chemical, biological, radiological, nuclear and explosive (CBRNE) sensors that detect, identify and assess CBRNE threats in real time. If a CBRNE event is detected, the modeling system will execute a real-time plume model, predict the impact on populations and recommend actions.



	Another distributed sensor network oriented toward detecting CBRNE threats is the Department of Homeland Security’s "Cell-All" program, which is currently in the research phase. The idea is to incorporate chemical, biological and radiological sensors into cell phones; upon detection of a possible threat, spatio-temporal information would be transmitted to emergency responders. Although there may be privacy concerns, such an approach would enable near-ubiquitous detection capability. Optical and hyperspectral remotely sensed imagery also provide useful information for environmental damage assessments associated with radiological disasters, particularly for civil and scientific users. Radionuclide contamination-induced vegetation stress can be discerned in reflectance spectra. In the event of a radiological disaster, such assessments can be critical for studying human exposure and ecosystem health, as well as for determining global food security, which was considered in the wake of the Chernobyl nuclear accident in 1986. Assessing and monitoring such contamination is also important for analyzing potential secondary contamination from, for example, fire events, which can transport contaminants via ash and smoke. In addition, unmanned aerial vehicles (UAVs) can be flown over areas of high radiological contamination without exposing personnel to dangerous radiation levels. Multiple UAVs may be required to serve this function, particularly in a time-sensitive disaster-response environment. In short, remotely evaluating vegetation damage in a hazards context is vital for agricultural production, and it can be used as a proxy indicator of environmental contamination from natural or man-made disasters. Widespread drought or fungal or insect infestation could threaten the ability of the increasing national and global human population to feed itself. As for contamination, some contaminants may persist for many years, potentially rendering large tracts of land unusable for human activity. These phenomena are clear threats to homeland security. Urban Damage Assessment An important dimension of disaster assessment in populated areas involves estimating the population affected by a given natural or technological event. In case of a disaster, a first priority is to assess and limit human injuries and fatalities; a second priority is to assess structural damage and recovery. Such assessments can be accomplished—at least in part—through remote sensing. However, a more accurate method entails the joint use of GIS-based modeling.



	A population database is required to estimate the population affected by a given disaster, and remote sensing can be an important tool for the task. Various approaches have been followed. For example, human population has been estimated at the local level by counting individual dwelling units. However, this approach is costly, time-consuming and impractical if a regional population base is to be compiled. Other approaches may be more appropriate. For example, an approach that employs land cover as a variable may be more accurate than some conventional methods, as land cover class is one of the best indicators of human population density. The LandScan project, developed at the Oak Ridge National Laboratory (ORNL), is a methodology that employs such information. Specifically, global population at a grid cell size of 1 kilometer is estimated primarily based on land cover, transportation network proximity, topography (slope) and high-resolution imagery. LandScan is being refined at ORNL to yield LandScan USA, a high-resolution database that includes nighttime/residential and daytime distributions, which is important given the temporal dimension of potential disasters. LandScan or LandScan USA population data can be intersected with measured or modeled disaster maps to determine the threatened population. In addition, a population database can be used to refine previous nuclear weapons-based blast casualty rate models. Some previous casualty rate models were based primarily on data from the Hiroshima and Nagasaki bombings during World War II. The models were used later to estimate casualties in U.S. cities following a nuclear attack. Remote sensing approaches also can be used to assess structural damage and contamination from various sources. Regardless of whether a disaster is natural or man-made (e.g., earthquake- or bomb-induced), building damage assessment is often conducted in situ, especially when interior structural damage is involved (e.g., the 1993 World Trade Center bombing). However, radar imaging can be used to image targets through building walls, thereby aiding post-disaster interior structural damage assessment and search-and-rescue efforts.



	Remote sensing is particularly useful when the building damage is outwardly apparent. For example, high-resolution pre- and post-Hurricane Katrina satellite images of New Orleans reveal hurricane-induced damage to individual structures and flooded areas. Additionally, close-range remote sensing—in the form of digital video or in situ spectroradiometer data acquisition—can be used with in situ mobile GIS data collection, as well as high-resolution airborne or spaceborne images. Remotely sensed images for disaster monitoring can be manually/visually interpreted or analyzed via computer algorithms. For manual photographic interpretation, large-scale color-infrared (CIR) photography usually is preferred over other photographic data for various urban disaster-assessment tasks, including determining the number and type of buildings damaged, analyzing damage to utilities, and evaluating access routes to and evacuation routes from a damaged area. Information extracted from high-resolution optical image data can vary as a function of image collection time and viewing angle. Shadows, solar illumination variability and geometric errors can minimize the effectiveness of automated damage-assessment algorithms, so manual/visual image interpretation is still one of the most commonly used methods for remote damage assessment. Automated image analysis/feature-recognition algorithms often are used to rapidly process a large volume of digital images. The type and characteristics of acquired data should be considered for various urban remote sensing tasks. For emergency scenarios, pre-disaster images should be acquired every 1 to 5 years, with a spatial resolution of 1 to 5 meters and spectral coverage in the panchromatic, visible and near-infrared (NIR). For post-disaster imagery, panchromatic or NIR high-resolution (≤ 0.25 to 2 meters) data should be collected as soon as possible. In the event of significant cloud cover, radar data can provide useful disaster information. For post-disaster quantitative damage assessments that involve buildings/housing, transportation and utilities, 0.25- to 1-meter panchromatic and NIR imagery should be obtained within a couple of days. Urban areas exhibit high spatial complexity, and spectral mixing typically occurs across all spatial scales. As a result, even high-resolution imagery may not lead to accurate urban cover or damage classifications. In addition, characterizing pre- or post-disaster urban areas using only pixel-based approaches can be problematic. For example, per-pixel classifiers can produce structural clutter—i.e., a salt-and-pepper effect—in the classified image. Given the high spectral diversity of materials and the high spatial frequency in urban areas, a region-based approach may compliment spectral-oriented processing for urban classification. Change detection-based damage assessment using 2.5-D/3-D building models also can be conducted, and such models can be constructed via a variety of methods. For example, airborne laser altimetry and ground plans can be used to construct a 3-D building model in which a 3-D Hough transform extracts planar surfaces from the point data clouds. Digital photogrammetry also can be used to generate urban digital surface models. Conventional stereo image-matching is used to determine corresponding pixels in overlapping images, and the resultant surface model entails elevations based on the visible surfaces of building tops or vegetation.

	Interferometric synthetic aperture radar (IfSAR) also can be used for building characterization, including identification and height estimation. IfSAR and hyperspectral data can be combined to produce urban 3-D surface feature geometry and topography. In addition, fused SAR and optical data can improve urban damage detection. Remotely Detecting Non-Radionuclide Contaminants Radionuclide-based contamination was discussed previously. However, remote sensing can be applied to other atmospheric plumes/contaminants from chemical and biological agents. For example, using a network of pre-positioned biochemical detectors and locating the point of origin of a chemical or biological weapon release makes it possible to predict the downwind plume movement with atmospheric nowcasting, which yields precise information regarding the current state of the atmosphere over a region. Data can be assimilated in real time, resulting in frequent updates. Urban-scale numerical models and detailed knowledge of physics and chemistry enable an effective operational response. It’s also possible to use aircraft-based data acquisition for detecting and monitoring airborne plumes. For example, the U.S. Environmental Protection Agency uses multispectral and hyperspectral passive infrared sensors to detect and classify absorption and emission features of chemical vapors. However, detection can be complicated if multiple chemical or biological agents mix together in the event of a terrorist attack. Spectral unmixing methods commonly employed in hyperspectral imaging can be used to classify and estimate the concentration of chemical and biological agents in a mixture. In addition, LiDAR—particularly Differential Absorption LiDAR—when used in conjunction with other sensors, may provide the ability to detect trace chemical concentrations. In addition, imaging spectroscopy can be used to evaluate disaster-associated contaminants. For example, following the 9/11 terrorist attack on the World Trade Center, the U.S. Geological Survey performed a laboratory reflectance spectroscopy analysis and used hyperspectral sensor data to assess and map the possible dissemination of potentially carcinogenic asbestiform dusts. Remote sensing also can be used to monitor other contaminants. For example, Hurricane Katrina-induced floodwaters in the New Orleans area were contaminated with oil and other constituents. SAR and optical approaches, for instance, have been used to detect and monitor oil spills for years. Ongoing Development Remote sensing enables a wide range of damage-assessment tasks. Likewise, a broad array of remote sensing technologies can be applied to a given problem, including photogrammetry, multispectral and hyperspectral imaging, SAR and LiDAR, among others. Sensors can be installed on airborne or spaceborne platforms, and in situ networks of fixed/discrete and dynamic/quasi-ubiquitous sensors play key roles. In addition, many related technologies and algorithms useful for damage assessment currently exist and have been proven in real-world scenarios. Although remote sensing isn’t a damage-assessment panacea, sensor and information-processing algorithms continue to evolve and improve. As a result, airborne technology and space-based imagery will be better situated to respond to a range of current and future threats.

See more Featured Articles See Featured Images Subscribe to Earth Imaging Journal Go to Home Page

[none]