A method is defined and tested for the characterization of agricultural land from multi-spectral imagery, based on singular value decomposition (SVD) and key vector analysis. The SVD technique, which bears a close resemblance to multivariate statistic techniques, has previously been successfully applied to problems of signal extraction for marine data and forestry species classification. In this study the SVD technique is used as a classifier for agricultural regions, using airborne Daedalus ATM data, with 1 m resolution. The specific region chosen is an experimental research farm in Bavaria, Germany. This farm has a large number of crops, within a very small region and hence is not amenable to existing techniques. There are a number of other significant factors which render existing techniques such as the maximum likelihood algorithm less suitable for this area. These include a very dynamic terrain and tessellated pattern soil differences, which together cause large variations in the growth characteristics of the crops. The SVD technique is applied to this data set using a multi-stage classification approach, removing unwanted land-cover classes one step at a time. Typical classification accuracy's for SVD are of the order of 85-100%. Preliminary results indicate that it is a fast and efficient classifier with the ability to differentiate between crop types such as wheat, rye, potatoes and clover. The results of characterizing 3 sub-classes of Winter Wheat are also shown.

This paper describes the capabilities of the NRSC's TSAR processing system, that provides a flexible approach to precise rectification, calibration and geo-referencing of satellite synthetic aperture radar (SAR) acquistions. An overview is given of how the product of TSAR enables the exploitation of ESA's ERS SAR images for land-use applications in mountainous and hilly terrain. In particular land-cover monitoring and geological analysis are considered.

Attempts to map different tropical forest types on a continental scale had varying success. Remotely sensed reflectance from forest canopies responds not only, or even primarily, to the species composition, but to a large number of factors, including soil type, forest structure, shadows and penology. This paper describes how detailed field data from sites in Bolivia are being used to develop and improve methods of tropical forest mapping from small scale satellite imagery. The paper presents preliminary work towards a novel procedure which characterizes canopy roughness from observations of forest structure and classifies imagery into canopy roughness classes by establishing a link between the shadow pattern information in the imagery and the canopy roughness of the forests. The inversion of the procedure will subsequently provide a means to extract forest structure information from satellite imagery and to generate improved maps of the tropical forest zone.

Initial assessment of the utility of ATSR-2 1 km data for retrieval of forest characteristics was made. Principal component analysis was applied to an image of a tropical forest site in order to assess the dimensionality of the 14 ATSR-2 features, which include 4 spectral and 3 thermal bands, obtained at 2 view angles. The combined dataset had a 5-dimensional structure. The optimal combination of spectral features for discriminating forest characteristics in this site were the four ATSR-2 nadir channels, plus either the forward red or green channel, which gave a 4-dimensional data structure. Sensitivity analysis of a forest BRDF model was employed to investigate the relationships between nadir and forward samples of the ATSR-2. Simulated relationships between nadir and forward samples, for stands of varying fractional ground cover and crown LAI, indicate weak non-linear relations for all spectral bands. BRDF simulation results indicate that some additional information, not contained in the nadir data, is contained in the forward data. A stochastic scene simulation, coupled with the BRDF model, was used to approximate the data structure of an ATSR-2 image. Results are interpreted in terms of a number of important factors operating in an ATSR-2 scene that may either reduce or enhance BRDF effects in the measured radiance, and thus affect the success of inverse BRDF algorithms. These factors include sensor spatial resolution, scales of variation of surface cover features, and differences in the sampling characteristics between nadir and forward scans. Further testing of ATSR-2 data, including the utility of its spectral bandwidths and its multi-temporal capability, is required for a variety of different forested sites.

The identification of plant communities in natural areas is fundamental to many applications. Among these are conservation and management of natural resources, production of thematic maps, and monitoring of relevant phenomena such as expansion of built-up areas and loss of grazing and cultivated lands. Multi-temporal radiometric characteristics as derived from ERS- 1 SAR and Landsat TM images have been analyzed for an area, located 30 km NW of Rome, which was studied by extensive ground surveys. The purpose of this study was to compare the information content of SAR images and Landsat TM bands analyzed together with topographical data, such as elevation, slope, and aspect, in order to develop an efficient classification methodology. Prior to classification, SAR images were corrected for the effect of scanning angle. Homogeneously vegetated areas found at different angle in the images allowed evaluation of the efficiency of this method. Canonical variable analysis was used to compare SAR and TM data in combination with topographic variables. After that, all the variables available were selected by a stepwise discriminant analysis for classification purposes. The variables considered in the model were used for the classification of test areas with an algorithm of linear discriminant analysis.

Assessment of ERS-1 SAR data for the monitoring of fallow land was performed over the Pointe Du Raz (Brittany, France), an agricultural region intensively crisscrossed by hedges delimiting small parcels. Comparison of the backscattering coefficients shows that from October to April, the identification of fallow land is possible while during late spring and summer identification appears more difficult. Texture analysis based on the texture spectrum approach does not allow the discrimination between waste land and cultivated land. SAR data could therefore be used for fallow land monitoring through a radiometric approach while the textural approach requires further investigation.

It is well known that reflectance of Earth surface largely depends upon amount of biomass, crop type, development stage, ground coverage. The knowledge of these parameters -- together with groundbased meteorological data -- allows for the estimate of crop water requirements and their spatial distribution. Recent research has shown the possibility of using multispectral satellite images in combination with other information for mapping crop coefficients in irrigated areas. This approach is based on the assumption that crop coefficients (K_c) are greatly influenced by canopy development and vegetation fractional ground cover; since these parameters directly affect the reflectance of cropped areas, it is possible to establish a correlation between multispectral measurements of canopies reflectance and the corresponding K_c values. Within this frame, two different approaches may be applied: (1) definition of spectral classes corresponding to different crop coefficient values and successive supervised classification for the derivation of crop coefficients maps; (2) use of analytical relationships between the surface reflectance and the corresponding values of vegetation parameters, i.e., the leaf area index, the albedo and the surface roughness, needed for the calculation of the potential evapotranspiration according to the combination type equation. The two different techniques are discussed with reference to the results of their application to specific case-studies. The aim of this report is to illustrate the suitability of remote sensing techniques as an operational tool for assessing crop water demand at regional scale.

Fluorescence lidar is an interesting tool for the control of vegetation status, monitoring stresses also at an early, previsual stage. However, a complete physiological interpretation of the fluorescence spectra and their variations related to the plant stresses is still missing. A research project dealing with both instrument development and data interpretation has been carried out at our institutes in the framework of the national programs IPRA and RAISA (technology for agriculture) and of the EUREKA/LASFLEUR (EU380) project. Laboratory experiments dealt with the interpretation of the physiological significance of blue-green and red vegetation fluorescence and with the analysis of the influence of stresses on the spectral behavior. The experiments were carried out both in spectroscopy laboratory, in greenhouses under controlled conditions (i.e., photosynthetic photon flux density, net photosynthesis, and different stresses) and, finally, in the field from ground based and flying platforms using the fluorescence lidar FLIDAR3.

Laboratory measurements of laser induced chlorophyll fluorescence kinetics (Kautsky effect) on dark-adapted vegetation targets (maize, pine-tree) have been performed with a lidar fluorosensor by superimposing probe pulses upon an actinic light. The collected induction curves (fast rise and slow decline) have been used to reveal the occurrence of stresses and the damage produced by a pine-tree parasite. A new two-pulse LIF (laser induced fluorescence) methodology has been investigated both theoretically and experimentally, in view of remotely monitoring the plant photosynthetic activity. This technique may yield information upon the in-vivo photosynthetic processes of plants, revealing a possible stress status (nutrients depletion, presence of herbicides, photoinhibition, etc.). The lidar apparatus used contains two laser sources in order to differentially measure the chlorophyll fluorescence by means of a laser pump-and-probe technique. In fact LIF signals in the red chlorophyll band 690 nm may provide in-vivo information upon photosynthesis process in high order plants and algae. Laser pump-and-probe experimental tests, with excitation 355 nm or 532 nm, already detect the presence of herbicides, and the effects of plant exposure to thermal stresses and to low levels of gaseous pollutants. Laser measured fluorescence yields (Y) have been found to be consistent with those obtained by an in-situ fluorimeter (PAM). With proper choices of experimental parameters (pump and probe laser intensities), Y approaches the theoretical value expected for a healthy dark-adapted plant.

It is known that the investigation of polarization structure of scattered optical radiation considerably enlarges the size of information on physical properties of objects interacting with radiation. For remote sensing the modulation Stokes-polarimeter was used. The main feature of this polarimeter is an electro-optical control converter of polarization in the receiving channel. It allows us to reach the time of single measurement up to a few microseconds and the accuracy does not exceed 0.1%. The measurement process is completely automatic. It carried out the investigation of the following objects: leaves of birch, oak, poplar; geranium; moss. The main consequences are: (1) polarization properties enable us to differ one kind of leaf from another; (2) the rear side of the leaf from the right one; (3) to monitor the degree of natural humidity; (4) to define whether the surface of a given foliage is humid (say, due to rain) or dry. All that and the theoretical investigations carried out by the author now enable us to conclude that the polarization characteristics are extremely informative and may be the basis of the process of recognition, monitoring, and classification of various kinds of foliage and grass covers.

The increase in accuracy of soil moisture retrieval from radar data requires a complete and accurate ground truth for calibration and verification. In order to support microwave backscattering models, one has to provide reliable permittivity and volumetric water content profiles of the soil. The conventional methods do not meet these requirements. The gravimetric sampling does not even deliver the volumetric water content. It has to be completed with an inaccurate measurement of the soil bulk density. Furthermore, the frequent probing of large areas is a tedious and expensive task. The electric conductivity is also a poor indicator due to its strong dependency on the salt content. Neutron probes have less and less acceptance and suffer from the shortcomings in measuring thin surface layers. The remaining possibility is the dielectric moisture determination by rf or microwaves, preferably in the optimal frequency range between 100 MHz and 1000 MHz. The measurement at single frequencies is better than the use of broadband pulses in the time domain, because of the capability to convert permittivities measured at one frequency to permittivities at the radar frequencies. Based on the measurement of the permittivity we have developed a combined sensor system for the ground truth. Our previously published frost-calibrated in-situ device evaluates the permittivity profile, the volumetric water content profile and the physical properties of the soil in regard to the permittivity of the solid component and in regard to the bound water. It uses a frequency of 250 MHz and it is installed at stationary locations. A second portable sensor extends the measuring area and completes the system. With frequencies of 500 MHz and 1000 MHz it is intermediate to the in-situ sensor and also to the radar frequencies in L-, C- and X-band. The scanning rate is about 100 m² in 10 minutes, which is suited for extensive field campaigns. This portable sensor is based on a Goubau-type open-ended surface waveguide, where a simple dielectric coated stick touches the soil surface. Its sensing area is sufficiently large to integrate over soil inhomogeneities. The diameter of the area and the penetration depth of the electromagnetic wave are adjusted by the frequency and the dimensions of the stick. The relation between permittivity and water content is determined by the in-situ sensor, resulting in an overall system accuracy of better than 6% for the volumetric water content.

This paper demonstrates that the soil physical properties can be estimated using temporal variation of surface soil moisture derived from remote sensing. Passive microwave remote sensing was employed to collect daily soil moisture data across the Little Washita watershed, Oklahoma, for the period between June 10-18, 1992. The ESTAR instrument operating at L band was flown on a NASA C-130 aircraft. Brightness temperature data collected at a ground resolution of 200 m were used to derive the spatial distribution of surface soil moisture. Analysis of temporal soil moisture information and soils data reveals a direct relationship between changes in soil moisture and soil texture. Areas identified by loam/silt loam soils are characterized by higher changes of total soil moisture and those of sand/sandy loam by remarkably lower amounts. Analysis suggests that two-day initial drainage of soil, measured from remote sensing, is related to saturated hydraulic conductivity (K_sat). A methodology has been developed to employ remotely sensed data for estimation of profile K_sat using a hydrologic model and a GIS. Model simulations have yielded good correlations between soil moisture change and K_sat. The results have potential applications to obtain quick estimates of spatial distributions of soil properties over large area for input to mesoscale hydrologic and global circulation models.

Simulated SPOT 4 data are assessed in environmentally sensitive wetland and regional agro- environmental policy monitoring. SPOT 4 was simulated via the fusion of SPOT XS and Landsat TM 5 SWIR bands. Using SERTIT's Region-Alsace GIS a result evaluation was performed on color composites, specific indices, tasseled cap (TM), PCA and Gram-Schmidt (SPOT4) indices calculations. The future SPOT 4 with its new MIR band and with less data provides more pertinent information than the existing Landsat TM and also than SPOT 3, systems on wetland environmental monitoring. This is particularly so when considering future cadastral level applications and its better spatial resolution.

In this paper we present some results of surface soil moisture retrieval using data from ERS- 1/SAR collected in 1994. Concurrent to ERS-1/SAR overflights field data were collected over two subcatchments of the Zwalm catchment in Belgium. After presenting a data processing procedure we apply a theoretical backscattering model to the data set. By comparing the calculated backscattering coefficients to the observed ones, we conclude that the measurements of surface roughness parameters in current remote sensing campaigns are not accurate enough for retrieving soil moisture using theoretical backscattering models. A method based on simultaneous retrieval of surface roughness parameters and soil moisture using multiple ERS- 1/SAR measurements is hence proposed. Promising results on retrieved soil moisture confirm the validity of the proposed method.

Microwave brightness temperatures were measured over a young wheat crop of varying canopy density during the First NOPEX Concentrated Field Effort near Uppsala, Sweden. Measurements were made with a portable ground-based microwave radiometer system at frequencies of 1.4 GHz and 5 GHz. The brightness temperature at each wavelength was measured at horizontal, vertical, and 45 degree polarization, and also at incidence angles ranging from nadir to 60 degrees. Four full measuring sequences were made; at full canopy, after each of two thinnings, and of the bare soil. A second series of measurements was also made over a similar plot which had been flooded with water to a depth of approximately 2 - 4 cm. Supporting data such as soil and canopy temperature, soil moisture, other soil physical properties, and canopy biomass were also collected. Preliminary results of some of the field measurements and comparisons with theoretical models are presented and discussed.

Nimbus/SMMR data have been used together with an extensive data base of soil moisture (1984-1987) to study the potential for large scale surface moisture monitoring. In order to study the influences of vegetation on the signals gathered from space, a radiative transfer model has been used to simulate the below canopy emissivity. Vegetation characteristics were subsequently estimated from inverse modeling and related to NDVI (synergistic approach) and to H- and V-polarization signatures (dual-polarization approach). Results of both approaches will be presented and compared. In order to validate the dual-polarization concept, which was applied to account for the radiative transfer properties of the vegetation, field measurements of single scattering albedo and optical depth were carried out recently for an agricultural crop.

The techniques of soil moisture and erosion state estimation using multichannel airborne radar remote sensing system are considered. The experimental data proving the efficiency of the proposed approach and showing the dependence of scattered signal intensity on soil characteristics are presented. The algorithms of radar image preprocessing are discussed.

First results of a ground radiometric campaign carried out in Antarctica during the last Italian scientific expedition in winter 1994/95 are here presented. The measurements were collected by means of a 4-channel radiometer, fitting with the MSS and first 4 TM spectral intervals. More than 100 measurements were taken in different sites, concerning three main categories: snow cover, continental ice, and marine ice. The investigation consisted in the collection of radiance information accompanied by observations of physical properties of the surfaces according to the standards by the International Glaciological Society. The reflectance values of different snow/ice covers indicate some interesting remarks to the use of these results as a method for the calibration of satellite imagery to a better understanding of this particular environment.

The seasonal changes of snow cover are delineated during the snowmelt period of a hydrological year by digital image processing of Landsat-4 and 5 Multispectral Scanner System (MSS) and Thematic Mapper (TM) data. The study area is the upper part of the Cordevole river basin located in the eastern part of the Italian Alps. Digital elevation model, slope and hill shading maps are generated and used in the study. Satellite images are classified into snow, mixed snow, and aper using a supervised maximum likelihood algorithm and estimated separately for three elevation zones of the catchment. The satellite derived information is integrated with meteorological and hydrological data in a snowmelt runoff model using pre-determined values of degree day factors, snow and rain runoff coefficients, and recession coefficient. The model performance evaluation indicated 0.94, 0.89, and +4.6 for correlation coefficient (r), Nash Sutcliffe coefficient (R²) and percentage volume deviation (D_v), respectively. The results are in good agreement when compared to those published for the World Meteorological Organization test basins. The relative advantages and disadvantages of using remote sensing data from various satellites for snowmelt runoff forecast are discussed.

Two ERS-1 SAR images of different dates were chosen to map flooding waters in the Camargue area of France. Comparative interpretation of the images is done principally by taking into account the nature of interaction between radar waves and water dipoles in the flooding waters, and its effect on the magnitude of retrodiffusion. Zones of low radiometric value are outlined on the image taken during the inundation period. In order to avoid artefacts and surfaces which are not covered by flooding waters, we took a systematically defined narrow radiometric interval that comprises inundated surfaces, and then we applied filters. Logical difference between the two superimposed scenes gave only the surfaces covered by flooding waters. As a result, flooding waters were mapped between Gimeaux and St. Gills, and around Lakes of Vaccares, and we put up a method of cartography of inundated zones from multidate ERS-1 SAR images on a flat area close to a sea. However, minor confusions due to muddy lakes and sediment saturated permanent water bodies remain possible.

In August 1990 some of the subwatersheds of the Real Collobrier research and experimental basin (Var-France) were totally or partially burnt by a forest fire. The concerned catchments provide an interesting case study of the hydrological consequences of a fire. Thus, since 1990, a research program has been developed to better understand the effect of vegetation on runoff genesis during the post-fire years. This program combines yearly vegetation surveys on experimental plots, remote sensing images (SPOT images) and hydrological information. The methodology developed is based on a comparative analysis of post-fire hydrological behaviors of subcatchments, and yearly mapping of post-fire vegetation recovery by using remote sensed data. This paper presents the results obtained after 4 years of observation (1990-1994). The effect of the vegetation destruction and recovery is clearly put in evidence. First attempts to analyze, in terms of process, the post-fire hydrological behaviors of the burnt subcatchments are presented.

The objectives of the presented work are on developing methods for extracting hydrologically relevant information from remote sensing data. Recognizing the difficulties in applying former land use and land cover classification systems for hydrological studies of river basins, a land use and land cover classification system for the Sauer River Basin, the Ruhr University Bochum-Land Use and Land Cover Classification System (RUB-LUCS), has been developed for providing distributed information for hydrological modeling, for detection of distributed land use changes and for study of scale problems in land use and land use change detection. Applying this system to Landsat data land use time series have been created for hydrological modeling of effects of man-made changes in the Sauer River Basin. To answer the question, what spatial resolution should be used in characterizing land use and land use changes in hydrological modeling of a mesoscale river basin, a study on scale problems has been conducted. As a result, a hydrological significant resolution is found as 1 km by km.

This paper presents a method of satellite remote sensing data analysis for comparative evaluation of inland waters optical quality, which would not necessarily require any in situ measurements. Remote sensing parameters were extracted by digital analysis of two Landsat Thematic Mapper images corresponding to different seasonal situations. The experiment was carried out on the Lake Garda (Italy), where distinct sub-basins can be individuated either for morphological stand-point or for their different water conditions. Subsurface irradiance reflectance values and chromaticity coordinates were derived from radiometric transects covering the different waters from the lake margin to the pelagic environment, and used to assess both spatial pattern and seasonal dynamics of the lake water quality.

A compact portable Fourier transform spectrometer is described. The simplicity of the instrument design and monolithic structure of the interferometer make it suitable for use in satellites or man-portable terrestrial measurements. Conceptual designs for hyperspectral imaging applications in the near-UV through near-IR, and in the mid-IR are presented. Imaging and spectral performance predictions are given using ray and beam tracing software to synthesize multispectral interferograms.

Dasa (Daimler-Benz Aerospace) currently is in charge of an ESA contract for the PRISM feasibility study, including MATRA MS/France and SIRA/UK. The main PRISM mission objectives are to monitor selected target areas with emphasis on bio-chemical stress analyses which are the subject of changes such as vegetation, in particular deforestation; further areas include glaciers, land/water boundaries, and similar features. The Dasa study team makes use of a similar ESA instrument currently the subject of dedicated subunits technology developments, the HRIS (high resolution imaging spectrometer), also lead by Dasa. Its main features and technology developments are covered in a survey. PRISM is covering most of the HRIS features, i.e., it will have an imaging spectrometer segment for the visible/NIR and the SWIR spectrum (450-2350 nm) as HRIS, but implements four MIR and TIR channels to provide surface temperature data, represented by linear detector arrays (essentially the features of another currently pursued ESA development program designated HRTIR). In a first study phase it also included a novel feature, a wide-angle two-axis pointing system for along- and across-track pointing in order to yield multi-directional reflectance data of dedicated targets, which turned out to be the greatest design driver in the instrument, it was omitted in the later phases. The geometric sampling interval is defined to 50 m (design goal) at 50 km swath. PRISM is currently studied in a feasibility study in a competitive way by two teams, one being lead by Dasa. At the time of the Symposium both studies are completed. The results of the Dasa team are presented. Since PRISM covers most of the HRIS mission it is rather likely that in the case of a successful concept feasibility proof (performance and technology wise), it may later replace the HRIS instrument. It is anticipated that a major part of the technology developments currently performed for HRIS will be utilized -- with some modifications -- for PRISM as well.

Since the 1980s the airborne imaging spectrometers have been more and more used for remote sensing of the environment. At the beginning they were utilized for geological investigations, whereas in the following years they have been applied to study vegetation, sea, atmosphere, snow and ice. Moreover that kind of instrumentation will be placed onboard different satellites. The Daedalus AA5000 Multispectral Infrared and Visible Imaging Spectrometer (MIVIS) is a very promising whisk-broom system, operating at 12 bits resolution in 102 narrow spectral bands from the visible to the thermal infrared. The MIVIS is flown onboard of a CASA 212 aircraft by the CNR-LARA Project. In the summer of 1994 a measurement campaign was carried out in South Italy and hyperspectral data were acquired during the overflight of the volcano Etna from inland to the sea. In order to evaluate the MIVIS performances, we developed some software and algorithms which can perform hyperspectral image processing and visualization under the UNIX operating system with the X Window System. By means of these tools the data were analyzed. Particularly, by applying the Fast Fourier Transform algorithm to selected spectral channels, qualitative information on noise patterns affecting MIVIS images was obtained. Preliminary results are shown and discussed.

The ISES model provides developers and users of hyperspectral optical instrumentation the means to evaluate instrument level performance as well as simulate hyperspectral images from space-based and airborne remote sensors. Instrument features modeled are: the foreoptics and monochromator optical design including the dispersive element, a large area two dimensional detector focal plane array, and the associated signal and data processing. The ISES code has been used to evaluate the performance of satellite based hyperspectral instrument design concepts as well as simulate space-based imagery collections using high spatial and spectral resolution airborne imagery data bases.

This paper describes a new approach to the calibration of space radiometers, building on developments made to improve primary terrestrial scales. The techniques described here apply to any space radiometer, whether it has imaging capability or not. For the purpose of this paper all space radiometers will be treated as a generic 'space filter radiometer' and may consist of detector arrays or single element detectors,interference filters or gratings, diffusers and imaging optics. The radiometer could also be operating in any part of the optical spectrum (UV, visible, or IR). To accurately determine the true spectral radiance of the Earth by the measurements made at the top of the atmosphere requires a precise knowledge of the radiometric transfer function of the atmosphere or a means of correcting for us using some form of vicarious calibration. This paper will not consider this aspect of calibration, although the techniques may help to improve some areas of vicarious calibration. The paper will instead concentrate on techniques to improve the accuracy of ground calibration of such instruments and on-board reference standards.

A multispectral intensified CCD imager combined with a ring laser gyroscope based inertial measurement unit was flown on the Space Shuttle Discovery from July 13-22, 1995 (Space Transport System Flight No. 70, STS-70). The camera includes a six position filter wheel, a third generation image intensifier, and a CCD camera. The camera is integrated with a laser gyroscope system that determines the ground position of the imagery to an accuracy of better than three nautical miles. The camera has two modes of operation; a panchromatic mode for high-magnification imaging [ground sample distance (GSD) of 4 m], or a multispectral mode consisting of six different user-selectable spectral ranges at reduced magnification (12 m GSD). This paper discusses the system hardware and technical trade-offs involved with camera optimization, and presents imagery observed during the shuttle mission.

PRISM (processes research by an imaging space mission) is a spaceborne hyperspectral imager studied by ESA for a future satellite devoted to land surface research. The prime objective of the mission would be the observation of biophysical processes from space at the local to regional scale. PRISM is designed to be flown on a medium-size satellite in a polar sun-synchronous orbit, providing coregistered spectral images in two spectral regions: the visible and short-wave infrared range (450 - 2350 nm), where it works as an imaging spectrometer with a spectral resolution of about 10 nm, and the thermal infrared range from 3.5 micrometer to 12.3 micrometer, divided in 4 spectral bands with a typical width of about 1 micrometer. The spatial resolution on ground is about 50 m and the swath about 50 km. Across-track pointing capability is available for convenient accessibility of the whole Earth. The objectives and requirements are described in this paper with a particular emphasis on the derivation of the radiometric requirements. The status of the instrument studies and of the related technology developments are also summarized.

Measurements of temporal reflectance signatures as a function of growing season for sand live oak (Quercus geminata), myrtle oak (Q. myrtifolia, and saw palmetto (Serenoa repens) were collected during a two year study period. Canopy level spectral reflectance signatures, as a function of 252 channels between 368 and 1115 nm, were collected using near nadir viewing geometry and a consistent sun illumination angle. Leaf level reflectance measurements were made in the laboratory using a halogen light source and an environmental optics chamber with a barium sulfate reflectance coating. Spectral measurements were related to several biophysical measurements utilizing optimal passive ambient correlation spectroscopy (OPACS) technique. Biophysical parameters included percent moisture, water potential (MPa), total chlorophyll, and total Kjeldahl nitrogen. Quantitative data processing techniques were used to determine optimal bands based on the utilization of a second order derivative or inflection estimator. An optical cleanup procedure was then employed that computes the double inflection ratio (DIR) spectra for all possible three band combinations normalized to the previously computed optimal bands. These results demonstrate a unique approach to the analysis of high spectral resolution reflectance signatures for estimation of several biophysical measures of plants at the leaf and canopy level from optimally selected bands or bandwidths.

An off-nadir canopy reflectance model, the Liang and Strahler algorithm for the coupled atmosphere and canopy (CAC) model, was used to simulate multi-angle reflectances based on various combinations of canopy biophysical parameters. Biophysical parameters such as leaf angle distribution and leaf area index were input to the CAC model along with reflectances of leaf, soil, and aerosol optical depth. The CAC model, however, can only be inverted through numerical iterations and it is extremely difficult to use for retrieval of those biophysical parameters with ordinary inversion methods. In order to retrieve those biophysical parameters, we employed an error back-propagation feed forward neural network program. We constructed a number of neural network models based on the simulated results from the CAC model. Ideally, through network training we would like to have a neural network model that uses the multi-angle reflectances as its inputs and output simultaneously all the biophysical parameters, the component reflectances of leaf and background soil, and the aerosol optical depth of the atmosphere. We have not yet reached this objective due to the requirement of an extremely large amount of calculation. In this paper, we report the results obtained from retrieving any individual parameter from multi-angle reflectances and results obtained from simultaneously retrieving some combinations of two parameters. We tested the use of a different number of multi-angle reflectances as input to the neural networks. This number varies in the range of 1 - 64. The test results show that a relative error between 1-5% or better is achievable for retrieving one parameter at a time or two parameters simultaneously

A technique has been developed to improve alpine-region spectral mixture analysis estimates of snow-covered area. Snow reflectance in near infrared wavelengths is sensitive to snow grain size while insensitive in visible wavelengths. Alpine regions often exhibit significant snow grain size gradients due to changes in aspect and elevation. A suite of snow image endmembers corresponding to the region's snow grain size range were extracted. Mixture models with fixed vegetation, rock, and shade were applied with each snow endmember to AVIRIS data collected over Mammoth Mountain, Calif., April 5, 1994. For each pixel, the snow-fraction estimated by the model with least mixing error (rms) was chosen to produce an optimal map of snow-covered area. Fraction under/overflow analysis and limited residuals analysis were performed on the test results.

The aim of the present work is to develop a new methodology suitable for landslide and flooded area extraction assessment using remotely sensed data acquired by satellite platform. This study was conducted after the still famous flood event that occurred in Piedmont (Italy) during November 1994. Caused by the heavy precipitation and by slope instability (due both to natural and artificial causes) in the southern part of Piedmont (and in particular in the Langhe area), an impressive series of catastrophic events occurred with the flooding of almost all of the area rivers and a rather large number of landslides in correspondence to the most hazardous slopes. During those days and in the following, several multispectral images were acquired by both airborne and satellite platform; these images are, still now, a not totally exploited source of data, suitable for the analysis of the territorial effects of such tremendous catastrophic phenomena. Particular attention has been paid to the setting up of a methodology suitable to establish the relationships between landslide (essentially debris flows), flooded areas and radiance and emittance data coming from the earth surface and recorded by the sensors mounted on board artificial satellites. All the data have been georeferenced, pre- processed, filtered and the images classified in order to evaluate from training samples a typical debris flow and water spectral signature.

The availability of MIVIS hyperspectral data, deriving from an aerial survey recently performed over a test-site in Lake Garda, Italy, gave the possibility of a preliminary new insight in the field of specific applications of remote sensing to shallow water analysis. The spectroradiometers in the visible and in the thermal infrared were explored in particular, accessing to helpful information for the detection of bio-physical indicators of water quality, either related to the surface/sub-surface of waters or to the bottom of the lake, since the study area presents very shallow waters, never exceeding a 6-meter depth in any case. Primary interest was the detection of man-induced activities along the margins, like sewage effect and sedimentary structure in the bottom or algal bloom. Secondly, a correlation between absorbivity coefficients in the visible bands and bathimetric contour lines in the proximity of the marginal zone of the lake was accomplished, by means of two indicative spectroradiometric transects.

On February 1994 a large area close to Trecate was affected by an oil blow-out from an AGIP rig located within the Ticino Regional Park. One month later an airborne survey has been carried out in the framework of the CNR Lara Project, by utilizing the Daedalus AA5000 MIVIS spectrometer with 102 channels from visible to thermal infrared. Different authors stress, for oil slicks discrimination, the utility of laser and microwaves based techniques, but the high spatial and spectral MIVIS resolutions can improve the detection of the relative coverage by spilled oil. This task has been performed by applying hyperspectral unmixing methods to the MIVIS calibrated data, obtaining an oil fractional image with respect to other chosen end-members. The analysis has shown a good agreement between the results of the unconstrained unmixing technique applied to MIVIS data and the ground truths, offering a tool useful to quantify in a synoptic overview the effects of oil spills over land, by relating the ppm of oil with the oil hyperspectral information gathered by MIVIS.

In recent years the improvement of the spectral and radiometric performances of optical detection systems has allowed the development of high spectral resolution remote sensors. Both passive and active sensors have taken advantage of this technological improvement and, as to the active ones, also from the contemporary development of new laser sources. At present, passive hyperspectral scanners with the simultaneous detection of more than 100 spectral images for each frame and fluorescence lidars with some hundreds of channels or with imaging capability are available. These instruments can now perform a remote spectroscopy of environmental targets with a spectral resolution similar to the laboratory one. However, in order to reduce the amount of data, it is necessary to exploit all the possible synergisms between active and passive optical sensors. This paper deals with the evaluation, the advantages and disadvantages of different solutions for the integration of active and passive sensors and formulates some hypothesis for sensor development and their use from flying platforms.

An approach to image enhancement based on information signal spectral abundance for passive imaging systems is proposed. The questions of spectral channel number selection and their parameter optimization are discussed taking into account different criterions of image quality. The peculiarities of the considered approach applications to radiometric and acoustic imaging systems are presented.

An instrument comprising the optical unit including the objective lens and acoustic-optical filter, detector, and electronic unit interfaced with the IBM PC AT has been designed and manufactured.