The availability of hyperspectral data has overcome the constraints and limitations of low spectral and spatial resolution imagery, and discreet spectral signatures. Hyperspectral images provide high spectral resolution data, with many narrow con¬tiguous spectral bands allowing for detailed applications.
1. Target detection, 2. Material mapping, 3. Material identification, 4. Mapping details of surface properties
In these cases, the additional information provided by hyperspectral imagery often provides results not possible with multispectral or other types of imagery. In target detection projects, investigators are generally trying to locate known target materials. This can sometimes involve distinguishing targets from very similar backgrounds, or locating examples of targets that are smaller than the nominal pixel size. For example, hyperspectral imagery has been used by military personnel to detect military vehicles under partial vegetation canopy, and to detect small military objects within relatively larger pixels.
The spectral characteristics of oil seeps and oil-impacted soils are generally too subtle to be detected by traditional multispectral sensors. In addition, oil seeps are limited in areal extent, and are usually mixed on the surface with other materials. Under these difficult conditions, hyperspectral sensors have sufficient spectral resolution to identify even small amounts of hydrocarbon- based material through their spectral signatures. In a material identification project, investigators do not know which materials are present in the scene. Under this scenario, the analysis is designed to use hyperspectral imagery for identifying the unknown materials. This analysis may also be accompanied by material mapping in which the identified materials are geographically located throughout the image. Material mapping is also performed with hyperspectral imagery when the materials present in the scene are known beforehand. For example, hyperspectral images have been used by geologists for mapping economically interesting minerals (e.g. Clark et al. 1995, 2003). They have also been used to map heavy metals and other toxic wastes within mine tailings in active and historic mining districts including superfund sites.
3.1. Mineral Diagnostic Spectra and Recognition
Mineral exploration using hyperspectral data is of increasing importance due to the fact that it provides an effective way (time and cost efficient) for analyzing sites of interest. There is an increasing need to effectively detect and quantify the presence of mineral signatures remotely, from airborne (or space based) sensors. The ability to do so provides geologists the opportunity to obtain geologic information well before they ever set foot in a particular landscape, significantly reducing costs and time associated with complex field explorations.
The ability to identify and map concentrations of key minerals can be used to rapidly and effectively survey large regions for new exploration targets. The technology is especially useful where the existing geology is not well understood or in areas of poor infrastructure and access. Identifying and mapping specific minerals associated with ore deposits greatly assists in the prioritization of exploration projects. In addition, with increasing world-wide need to ensure environmental compliance during the development of natural exploration using hyperspectral data is of increasing importance due to the fact that it provides an effective way (time and cost efficient) for analyzing sites of interest. There is an increasing need to effectively detect and quantify the presence of mineral signatures remotely, from airborne (or space based) sensors. The ability to do so provides geologists the opportunity to obtain geologic information well before they ever set foot in a particular landscape, significantly reducing costs and time associated with complex field explorations.
3.2. Hyperspectral Imaging Offers Huge Potential to Agriculture
Hyperspectral images acquired by aircraft and satellites have the potential to detect crop stress and diagnose its cause long before a farmer can spot the problem in his field with the naked eye. Farmers have heard the remote sensing oversell before, so let’s make one point clear: Hyperspectral imaging will not revolutionize precision farming overnight. But it is a technology that will have a major impact. Much research still must be completed, and the practical evolution of hyperspectral imaging in agriculture will occur much faster than it did with multispectral imagery. For this reason, some precision farming practitioners are already gearing up to handle the new data, and they will be the first to benefit from it.
Numerous factors are driving the rapid development of hyperspectral imaging applications. For example, NASA’s Lewis satellite will acquire 384 bands of data which will be provided at no charge to anyone who wants it. Unrestricted distribution will enable precision farming researchers to explore the data’s potential at relatively small expense. More importantly, many agriculture researchers are already using hyperspectral airborne imagery today. Many systems, such as the Jet Propulsion Laboratory’s AVIRIS and ITRES Corp.’s casi sensors, are available for private and public sector use. Increased competition is driving down the cost to purchase or lease these instruments.
New satellites with one-meter spatial resolution in a single band (i.e. panchromatic) will also enhance the utility of hyperspectral and multispectral imagery. Media publicity portrays these high-resolution satellites as the remote sensing tool for precision farming. This is an exaggeration. Spatial resolution offers great value in monitoring crop appearance, but it is spectral data that reveals the most information about plant stress and health. Fig.2 showing the CASI FCC image for study of agricultural farm management
Fig.2. CASI image showing the agricultural farm management (Source: Chung-hsin Juan, et.al 2000)
3.3. Detection of water quality
Remote sensing technology has been widely used in water resource applications (Gitelson and Merzlyak, 1996; Zagolski et al., 1996; Asner, 1998; McGwire et al., 2000; Stone et al., 2001; Coops et al., 2002; Underwood et al., 2003) and in particular hyperspectral remote sensing is emerging as the more in-depth means of investigating spatial, spectral and temporal varia¬tions in order to derive more accurate estimates of information required for water resource applications. This section briefly highlights applications of hyperspectral remote sensing in water resources, and is followed by a detailed review of the methods and applications of land- use and vegetation classification.
Conditions and parameters is one of the major advantages of hyperspectral remote sensing tech¬nologies. Hyperspectral reflectance has been widely used to assess water quality conditions of many open water aquatic ecosystems. This includes classifying the trophic status of lakes (Koponen et al., 2002; Thiemann and Kaufmann, 2002) and estuaries (Froidefond et al., 2002) characterizing algal blooms (Stumpf, 2001) and assessment of ammonia dynamics for wet¬land treatments (Tilley et al., 2003). Predictors of total ammonia concentrations using remotely sensed hyperspectral signatures of macrophytes in order to monitor changes in wetland water quality were developed by Tilley et al. (2003). Hyperspectral spectrometers have also proved useful in determining the total suspended matter, chlorophyll content (Hakvoort et al., 2002; Vos et al., 2003) and total phosphorus (Koponen et al., 2002). Much research has been undertaken in the estimation of chloro¬phyll content from remotely sensed images which is then used as an estimate for monitoring algal content and hence water quality. Since wavelengths corresponding with the peak reflectance of blue-green and green algae are close together it is more difficult to differentiate between them. However, hyperspectral imagers allow for improved detection of chlorophyll and hence algae, due to the narrow spectral bands which are acquired between 450 nm and 600 nm. (Hakvoort et al., 2002). Estimation and mapping of water quality constituents such as concentrations of dissolved organic matter, chlorophyll or total suspended matter from optical remote sensing technologies have proved to be use¬ful and successful and are being investigated for operational use (Hakvoort et al., 2002).
3.4. Flood detection and monitoring
These are constrained by the ina¬bility to obtain timely information of water conditions from ground measurements and airborne observations at sufficient temporal and spatial resolutions. Satellite remote sensing allows for timely investigation of areas of large regional extent and provides frequent imaging of the region of interest (Felipe et al., 2006). Until recently, near real-time flood detection was not possible, but with sensors such as Hyperion on board the EO-1 satellite this has been vastly improved (Felipe et al., 2006). According to research conducted by Felipe et al. (2006) auto¬mated spacecraft technology reduced the time to detect and react to flood events to a few hours. Advances in remote sensing, have resulted in the investigation of early warning systems with potential global applications. Most recent studies from NASA and the US Geological Survey are utilising satellite observations of rainfall, rivers and surface topography into early warning sys¬tems sys¬tems (Brakenridge et al., 2006). The estimation of discharge and flood hydrographs from hydraulic information obtained from remotely sensed data was assessed by Roux and Dartus (2006). Remote sensed images as used to estimate the hydraulic charac¬teristics which are then applied in routing modules to generate a flood wave in a synthetic river channel. Optimisation methods are used to minimise discrepancies between simulations and observations of flood extent fields to estimate river discharge (Roux and Dartus, 2006).
3.5. Measures of plant physiology and structure
Traditional methods for landscape-scale vegetation mapping require expensive, time-intensive field surveys. Remotely sensed data for the classification and mapping of vegetation provide a detailed accurate product in a time- and cost-effec¬tive manner. The availability of satellite and airborne hyper¬spectral data with its increased spatial and more critically fine spectral resolution offers an enhanced potential for the classifi¬cation and mapping of land use and vegetation. Due to the large number of wavebands, image processing is able to capitalise on both the biochemical and structural properties of vegetation (Underwood et al., 2003). The need for exploring these spectral properties is particularly important when we consider the limi¬tations of using traditionally available wavebands, where most of the land cover is grouped and identification of individual species is difficult.
These applications investigate the spectral reflect¬ance properties of plants, identifying key spectral wavebands related to specific plant physiological and structural character¬istics, hence deriving sensitive vegetation spectral indices for their non-destructive estimation. Remote sensing data have been exploited to estimate canopy characteristics by using empirical approaches based on spectral Analysis of hyperspectral remote sensing data has been carried out to estimate LAI for agricultural crops and forests. The accurate estimation of plant water status and plant water stress is essential to the integration of remote sensing into precision agricultural and forestry management. The potential to spectrally estimate plant physiological properties over relatively large areas, and to predict plant water status and plant water stress was demonstrated by Champagne et al., 2003 for agri¬cultural crops; and Stimson et al., 2005 and Eitel al., 2006 for forestry species. Their results indicate the potential use of veg¬etation spectral indices derived from various scales of remote sensing data for determining plant physiological properties and characteristics. These studies amongst others clearly indicate the improved estimates of plant physiological and structural characteristics from hyperspectral data, allowing for much more detailed spectral analyses and hence more accurate estimates.
3.6. Wetland mapping
It has gained increased recognition for the abil¬ity to improve quality of ecosystems. Sustainable management of any ecosystem requires, among other information, a thorough understanding of vegeta¬tion species distribution. Hyperspectral imagery has been used to remotely delineate wetland areas and classify hydrophytic vegetation characteristics of these ecosystems (Schmidt and Skidmore, 2003; Becker et al., 2005). The hyperspectral analysis identified key regions of the electromagnetic spectrum which provided detailed information for discriminating between and identifying different wetland species (Schmidt and Skidmore, 2003). Becker et al. (2005) performed a similar study based on coastal wetland plant communities which are spatially complex and heterogeneous. This study also emphasised the impor¬tance of hyperspectral imagery for identifying and differentiat¬ing vegetation spectral properties from narrow spectral bands focussing on the visible and near-infrared regions (Becker et al., 2005). A number of studies have investigated the potential of providing timely data for mapping and monitoring submerged aquatic vegetation which has been identified as one of the most important aspects of ecosystem restoration and reconstruction (Lin and Liquan, 2006). Such species have been termed ecologi¬cal engineering species and the quantification of their coverage and spectral reflectance properties is currently being researched (Lin and Liquan, 2006).
3.7. Pollutant Detection in Drainage System
The work carried out in Pennsylvania to determine the presence of a suspect pollutant in a drainage system. The test data was evaluated to determine if it was possible to construct a detection system for locating pollutants in and along waterways. The raw RGB filtered image is shown in fig.2 and processed image is shown in Fig.4.finally polluted drainage identification is shown in fig.5
By: M.RAJAMANICKAM
About the Author:
1. Target detection, 2. Material mapping, 3. Material identification, 4. Mapping details of surface properties
In these cases, the additional information provided by hyperspectral imagery often provides results not possible with multispectral or other types of imagery. In target detection projects, investigators are generally trying to locate known target materials. This can sometimes involve distinguishing targets from very similar backgrounds, or locating examples of targets that are smaller than the nominal pixel size. For example, hyperspectral imagery has been used by military personnel to detect military vehicles under partial vegetation canopy, and to detect small military objects within relatively larger pixels.
The spectral characteristics of oil seeps and oil-impacted soils are generally too subtle to be detected by traditional multispectral sensors. In addition, oil seeps are limited in areal extent, and are usually mixed on the surface with other materials. Under these difficult conditions, hyperspectral sensors have sufficient spectral resolution to identify even small amounts of hydrocarbon- based material through their spectral signatures. In a material identification project, investigators do not know which materials are present in the scene. Under this scenario, the analysis is designed to use hyperspectral imagery for identifying the unknown materials. This analysis may also be accompanied by material mapping in which the identified materials are geographically located throughout the image. Material mapping is also performed with hyperspectral imagery when the materials present in the scene are known beforehand. For example, hyperspectral images have been used by geologists for mapping economically interesting minerals (e.g. Clark et al. 1995, 2003). They have also been used to map heavy metals and other toxic wastes within mine tailings in active and historic mining districts including superfund sites.
3.1. Mineral Diagnostic Spectra and Recognition
Mineral exploration using hyperspectral data is of increasing importance due to the fact that it provides an effective way (time and cost efficient) for analyzing sites of interest. There is an increasing need to effectively detect and quantify the presence of mineral signatures remotely, from airborne (or space based) sensors. The ability to do so provides geologists the opportunity to obtain geologic information well before they ever set foot in a particular landscape, significantly reducing costs and time associated with complex field explorations.
The ability to identify and map concentrations of key minerals can be used to rapidly and effectively survey large regions for new exploration targets. The technology is especially useful where the existing geology is not well understood or in areas of poor infrastructure and access. Identifying and mapping specific minerals associated with ore deposits greatly assists in the prioritization of exploration projects. In addition, with increasing world-wide need to ensure environmental compliance during the development of natural exploration using hyperspectral data is of increasing importance due to the fact that it provides an effective way (time and cost efficient) for analyzing sites of interest. There is an increasing need to effectively detect and quantify the presence of mineral signatures remotely, from airborne (or space based) sensors. The ability to do so provides geologists the opportunity to obtain geologic information well before they ever set foot in a particular landscape, significantly reducing costs and time associated with complex field explorations.
3.2. Hyperspectral Imaging Offers Huge Potential to Agriculture
Hyperspectral images acquired by aircraft and satellites have the potential to detect crop stress and diagnose its cause long before a farmer can spot the problem in his field with the naked eye. Farmers have heard the remote sensing oversell before, so let’s make one point clear: Hyperspectral imaging will not revolutionize precision farming overnight. But it is a technology that will have a major impact. Much research still must be completed, and the practical evolution of hyperspectral imaging in agriculture will occur much faster than it did with multispectral imagery. For this reason, some precision farming practitioners are already gearing up to handle the new data, and they will be the first to benefit from it.
Numerous factors are driving the rapid development of hyperspectral imaging applications. For example, NASA’s Lewis satellite will acquire 384 bands of data which will be provided at no charge to anyone who wants it. Unrestricted distribution will enable precision farming researchers to explore the data’s potential at relatively small expense. More importantly, many agriculture researchers are already using hyperspectral airborne imagery today. Many systems, such as the Jet Propulsion Laboratory’s AVIRIS and ITRES Corp.’s casi sensors, are available for private and public sector use. Increased competition is driving down the cost to purchase or lease these instruments.
New satellites with one-meter spatial resolution in a single band (i.e. panchromatic) will also enhance the utility of hyperspectral and multispectral imagery. Media publicity portrays these high-resolution satellites as the remote sensing tool for precision farming. This is an exaggeration. Spatial resolution offers great value in monitoring crop appearance, but it is spectral data that reveals the most information about plant stress and health. Fig.2 showing the CASI FCC image for study of agricultural farm management
Fig.2. CASI image showing the agricultural farm management (Source: Chung-hsin Juan, et.al 2000)
3.3. Detection of water quality
Remote sensing technology has been widely used in water resource applications (Gitelson and Merzlyak, 1996; Zagolski et al., 1996; Asner, 1998; McGwire et al., 2000; Stone et al., 2001; Coops et al., 2002; Underwood et al., 2003) and in particular hyperspectral remote sensing is emerging as the more in-depth means of investigating spatial, spectral and temporal varia¬tions in order to derive more accurate estimates of information required for water resource applications. This section briefly highlights applications of hyperspectral remote sensing in water resources, and is followed by a detailed review of the methods and applications of land- use and vegetation classification.
Conditions and parameters is one of the major advantages of hyperspectral remote sensing tech¬nologies. Hyperspectral reflectance has been widely used to assess water quality conditions of many open water aquatic ecosystems. This includes classifying the trophic status of lakes (Koponen et al., 2002; Thiemann and Kaufmann, 2002) and estuaries (Froidefond et al., 2002) characterizing algal blooms (Stumpf, 2001) and assessment of ammonia dynamics for wet¬land treatments (Tilley et al., 2003). Predictors of total ammonia concentrations using remotely sensed hyperspectral signatures of macrophytes in order to monitor changes in wetland water quality were developed by Tilley et al. (2003). Hyperspectral spectrometers have also proved useful in determining the total suspended matter, chlorophyll content (Hakvoort et al., 2002; Vos et al., 2003) and total phosphorus (Koponen et al., 2002). Much research has been undertaken in the estimation of chloro¬phyll content from remotely sensed images which is then used as an estimate for monitoring algal content and hence water quality. Since wavelengths corresponding with the peak reflectance of blue-green and green algae are close together it is more difficult to differentiate between them. However, hyperspectral imagers allow for improved detection of chlorophyll and hence algae, due to the narrow spectral bands which are acquired between 450 nm and 600 nm. (Hakvoort et al., 2002). Estimation and mapping of water quality constituents such as concentrations of dissolved organic matter, chlorophyll or total suspended matter from optical remote sensing technologies have proved to be use¬ful and successful and are being investigated for operational use (Hakvoort et al., 2002).
3.4. Flood detection and monitoring
These are constrained by the ina¬bility to obtain timely information of water conditions from ground measurements and airborne observations at sufficient temporal and spatial resolutions. Satellite remote sensing allows for timely investigation of areas of large regional extent and provides frequent imaging of the region of interest (Felipe et al., 2006). Until recently, near real-time flood detection was not possible, but with sensors such as Hyperion on board the EO-1 satellite this has been vastly improved (Felipe et al., 2006). According to research conducted by Felipe et al. (2006) auto¬mated spacecraft technology reduced the time to detect and react to flood events to a few hours. Advances in remote sensing, have resulted in the investigation of early warning systems with potential global applications. Most recent studies from NASA and the US Geological Survey are utilising satellite observations of rainfall, rivers and surface topography into early warning sys¬tems sys¬tems (Brakenridge et al., 2006). The estimation of discharge and flood hydrographs from hydraulic information obtained from remotely sensed data was assessed by Roux and Dartus (2006). Remote sensed images as used to estimate the hydraulic charac¬teristics which are then applied in routing modules to generate a flood wave in a synthetic river channel. Optimisation methods are used to minimise discrepancies between simulations and observations of flood extent fields to estimate river discharge (Roux and Dartus, 2006).
3.5. Measures of plant physiology and structure
Traditional methods for landscape-scale vegetation mapping require expensive, time-intensive field surveys. Remotely sensed data for the classification and mapping of vegetation provide a detailed accurate product in a time- and cost-effec¬tive manner. The availability of satellite and airborne hyper¬spectral data with its increased spatial and more critically fine spectral resolution offers an enhanced potential for the classifi¬cation and mapping of land use and vegetation. Due to the large number of wavebands, image processing is able to capitalise on both the biochemical and structural properties of vegetation (Underwood et al., 2003). The need for exploring these spectral properties is particularly important when we consider the limi¬tations of using traditionally available wavebands, where most of the land cover is grouped and identification of individual species is difficult.
These applications investigate the spectral reflect¬ance properties of plants, identifying key spectral wavebands related to specific plant physiological and structural character¬istics, hence deriving sensitive vegetation spectral indices for their non-destructive estimation. Remote sensing data have been exploited to estimate canopy characteristics by using empirical approaches based on spectral Analysis of hyperspectral remote sensing data has been carried out to estimate LAI for agricultural crops and forests. The accurate estimation of plant water status and plant water stress is essential to the integration of remote sensing into precision agricultural and forestry management. The potential to spectrally estimate plant physiological properties over relatively large areas, and to predict plant water status and plant water stress was demonstrated by Champagne et al., 2003 for agri¬cultural crops; and Stimson et al., 2005 and Eitel al., 2006 for forestry species. Their results indicate the potential use of veg¬etation spectral indices derived from various scales of remote sensing data for determining plant physiological properties and characteristics. These studies amongst others clearly indicate the improved estimates of plant physiological and structural characteristics from hyperspectral data, allowing for much more detailed spectral analyses and hence more accurate estimates.
3.6. Wetland mapping
It has gained increased recognition for the abil¬ity to improve quality of ecosystems. Sustainable management of any ecosystem requires, among other information, a thorough understanding of vegeta¬tion species distribution. Hyperspectral imagery has been used to remotely delineate wetland areas and classify hydrophytic vegetation characteristics of these ecosystems (Schmidt and Skidmore, 2003; Becker et al., 2005). The hyperspectral analysis identified key regions of the electromagnetic spectrum which provided detailed information for discriminating between and identifying different wetland species (Schmidt and Skidmore, 2003). Becker et al. (2005) performed a similar study based on coastal wetland plant communities which are spatially complex and heterogeneous. This study also emphasised the impor¬tance of hyperspectral imagery for identifying and differentiat¬ing vegetation spectral properties from narrow spectral bands focussing on the visible and near-infrared regions (Becker et al., 2005). A number of studies have investigated the potential of providing timely data for mapping and monitoring submerged aquatic vegetation which has been identified as one of the most important aspects of ecosystem restoration and reconstruction (Lin and Liquan, 2006). Such species have been termed ecologi¬cal engineering species and the quantification of their coverage and spectral reflectance properties is currently being researched (Lin and Liquan, 2006).
3.7. Pollutant Detection in Drainage System
The work carried out in Pennsylvania to determine the presence of a suspect pollutant in a drainage system. The test data was evaluated to determine if it was possible to construct a detection system for locating pollutants in and along waterways. The raw RGB filtered image is shown in fig.2 and processed image is shown in Fig.4.finally polluted drainage identification is shown in fig.5
By: M.RAJAMANICKAM
About the Author:
MSC, PHD (REMOTE SENSING AND GIS)
