Search

The product generation from hyperspectral sensor data has high requirements on the processing infrastructure, both hardware and software. The Airborne Prism Experiment (APEX) processing and archiving facility has been set up to provide for the automated generation of level-1 calibrated data and user-configurable on-demand product generation for higher processing levels. The system offers full reproducibility of user orders and processing parameters by employing a relational database. The flexible workflow software allows for the quick integration of novel algorithms or the definition of new processing sequences. Reprocessing of data is supported by the archiving approach. Configuration management based on the database enables the control over different versions of processing modules to be applied. The system is described with a focus on the APEX instrument; however, its generic design allows adaptation to other sensor systems.

The airborne ESA-APEX (Airborne Prism Experiment) hyperspectral mission simulator is described with its distinct specifications to provide high quality remote sensing data. The concept of an automatic calibration, performed in the Calibration Home Base (CHB) by using the Control Test Master (CTM), the In-Flight Calibration facility (IFC), quality flagging (QF) and specific processing in a dedicated Processing and Archiving Facility (PAF), and vicarious calibration experiments are presented. A preview on major applications and the corresponding development efforts to provide scientific data products up to level 2/3 to the user is presented for limnology, vegetation, aerosols, general classification routines and rapid mapping tasks. BRDF (Bidirectional Reflectance Distribution Function) issues are discussed and the spectral database SPECCHIO (Spectral Input/Output) introduced. The optical performance as well as the dedicated software utilities make APEX a state-of-the-art hyperspectral sensor, capable of (a) satisfying the needs of several research communities and (b) helping the understanding of the Earth’s complex mechanisms.

Efficient and accurate imaging spectroscopy data processing asks for perfectly consistent (i.e., ideally uniform) data in both the spectral and spatial dimensions. However, real pushbroom-type imaging spectrometers are affected by various point spread function (PSF) nonuniformity artifacts. First, individual pixels or lines may be missing in the raw data due to bad pixels originating from the detector, readout errors, or even electronic failures. Second, so-called smile and keystone optical aberrations are inherent to imaging spectrometers. Appropriate resampling strategies are required for the preprocessing of such data if emphasis is put on spatial PSF uniformity. So far, nearest neighbor interpolations have been often recommended and used for resampling. This paper shall analyze the radiometric effects if linear interpolation is used to optimize the spatial PSF uniformity. For modeling interpolation effects, an extensive library of measured surface reflectance spectra as well as real imaging spectroscopy data over various land cover types are used. The real measurements are systematically replaced by interpolated values, and the deviation between original and resampled spectra is taken as a quality measure. The effects of nearest neighbor resampling and linear interpolation methods are compared. It is found that linear interpolation methods lead to average radiometric errors below 2% for the correction of spatial PSF nonuniformity in the subpixel domain, whereas the replacement of missing pixels leads to average errors in the range of 10%-20%

The organised storage of spectral data described by metadata is important for long-term use and data sharing with other scientists. Metadata describing the sampling environment, geometry and measurement process serves to evaluate the suitability of existing data sets for new applications. There is a need for spectral databases that serve as repositories for spectral field campaign and reference signatures, including appropriate metadata parameters. Such systems must be (a) highly automated in order to encourage users entering their spectral data collections and (b) provide flexible data retrieval mechanisms based on subspace projections in metadata spaces. The recently redesigned SPECCHIO system stores spectral and metadata in a relational database based on a non-redundant data model and offers efficient data import, automated metadata generation, editing and retrieval via a Java application. RSL is disseminating the database and software to the remote sensing community in order to foster the use and further development of spectral databases.

A physically based algorithm is used for automatic processing of MERIS level 1B full resolution data. The algorithm is originally used with input variables for optimization with different sensors (i.e. channel recalibration and weighting), aquatic regions (i.e. specific inherent optical properties) or atmospheric conditions (i.e. aerosol models). For operational use, however, a lake-specific parameterization is required, representing an approximation of the spatio-temporal variation in atmospheric and hydrooptic conditions, and accounting for sensor properties. The algorithm performs atmospheric correction with a LUT for at-sensor radiance, and a downhill simplex inversion of chl-a, sm and y from subsurface irradiance reflectance. These outputs are enhanced by a selective filter, which makes use of the retrieval residuals. Regular chl-a sampling measurements by the Lake’s protection authority coinciding with MERIS acquisitions were used for parameterization, training and validation.

The increasing quantity and sophistication of imaging spectroscopy applications have led to a higher demand on the quality of Earth observation data products. In particular, it is desired that data products be as consistent as possible (i.e., ideally uniform) in both spectral and spatial dimensions. Yet, data acquired from real (e.g., pushbroom) imaging spectrometers are adversely affected by various categories of artifacts and aberrations including as follows: singular and linear (e.g., bad pixels and missing lines), area (e.g., optical aberrations), and stability and degradation defects. Typically, the consumer of such data products is not aware of the magnitude of such inherent data uncertainties even as more uncertainty is introduced during higher level processing for any particular application. In this paper, it is shown that the impact of imaging spectrometry data product imperfections in currently available data products has an inherent uncertainty of 10%, even though worst case scenarios were excluded, state-of-the-art corrections were applied, and radiometric calibration uncertainties were excluded. Thereafter, it is demonstrated how this error can be reduced (

A physically based water constituent retrieval algorithm is used for the automatic processing of MERIS level 1B full resolution data. The algorithm requires several input variables for individual optimization with different sensors (i. e. channel calibration and weighting), aquatic regions (i. e. specific inherent optical properties) or atmospheric conditions (i. e. Aerosol models). The optical properties are derived from optical in situ measurements during concurrent MERIS data acquisition on 20 April 2007. Remaining parameters are iteratively optimized for best performance with 21 MERIS datasets of Lake Constance in the years 2003-2005, and validated with 11 datasets in 2006. Operational water quality sampling measurements acquired by local authorities serve as reference.

This study explores performance requirements for the retrieval of the atmospheric aerosol optical depth (AOD) by airborne optical remote sensing instruments. Independent of any retrieval techniques, the calculated AOD retrieval requirements are compared with the expected performance parameters of the upcoming hyperspectral sensor APEX at the reference wavelength of 550nm. The AOD accuracy requirements are defined to be capable of resolving transmittance differences of 0.01 to 0.04 according to the demands of atmospheric corrections for remote sensing applications. For the purposes of this analysis, the signal at the sensor level is simulated by radiation transfer equations. The resulting radiances are translated into the AOD retrieval sensitivity (ΔÄλaer ) and compared to the available measuring sensitivity of the sensor (NE ΔLλsensor). This is done for multiple signal-to-noise ratios (SNR) and surface reflectance values. It is shown that an SNR of 100 is adequate for AOD retrieval at 550nm under typical remote sensing conditions and a surface reflectance of 10% or less. Such dark surfaces require the lowest SNR values and therefore offer the best sensitivity for measuring AOD. Brighter surfaces with up to 30% reflectance require an SNR of around 300. It is shown that AOD retrieval for targets above 50% surface reflectance is more problematic with the current sensor performance as it may require an SNR larger than 1000. In general, feasibility is proven for the analyzed cases under simulated conditions.

Hyperspectral imaging (HSI) sensors suffer from spatial misregistration, an artifact that prevents the accurate acquisition of the spectra. Physical considerations let us assume that the influence of the spatial misregistration on the acquired data depends both on the wavelength and on the across-track position. A scene-based method, based on edge detection, is therefore proposed. Such a procedure measures the variation on the spatial location of an edge between its various monochromatic projections, giving an estimation for spatial misregistration, and also allowing identification of misalignments. The method has been applied to several hyperspectral sensors, either prism, or grating-based designs. The results confirm the dependence assumptions on lambda and theta, spectral wavelength and across-track pixel, respectively. Suggestions are also given to correct for spatial misregistration.

This paper addresses the main goals and objectives of the Hyperspectral Imaging Network (HYPER-I-NET), a recently started Marie Curie Research Training Network. The project is designed to build an interdisciplinary research community focusing on hyperspectral imaging activities. The core strategy of the network is to create a powerful interdisciplinary synergy between different domains of expertise closely related to hyperspectral imaging activities in Europe, ranging from sensor design and flight operation to data collection, processing, interpretation, and dissemination. Our main goals in this paper are to present the project to the Geoscience and Remote Sensing community and to provide an overview of the planned activities in each sub-activity covered by the network.

In the scope of hyperspectral airborne imaging spectrometer (APEX) design activities, the acceptable sensitivity of linear polarization of the spectrometer is analyzed by assessing the amount of polarization of reflected light in the atmosphere-surface system. A large number of calculations is performed for a wide variaty of viewing geometries to study the influences of aerosol models, natural surfaces and flight altitudes over the spectral range from the near-UV to the short-wave infrared (SWIR). Thereinafter the design of the imaging spectrometer is outlined accounting for these requirements and a method of partially correcting the instrument polarization sensitivity is briefly introduced. APEX design and post-processing capabilities will enable to reduce the influence of polarization sensitivity of at-sensor radiance and its higher-level products generated for most of the observation conditions.

Spectro-directional surface measurements can either be performed in the field or within a laboratory setup. Laboratory measurements have the advantage of constant illumination and neglectable atmospheric disturbances. On the other hand, artificial light sources are usually less parallel and less homogeneous than the clear sky solar illumination. To account for these differences and for determining for which targets a replacement of field by laboratory experiments is indeed feasible, a quantitative comparison is a prerequisite. Currently, there exists no systematic comparison of field and laboratory measurements using the same targets. In this study we concentrate on the difference in spectro-directional field and laboratory data of the same target due to diffuse illumination. The field data were corrected for diffuse illumination following the proposed procedure by Martonchik . Spectro-directional data were obtained with a GER3700 spectroradiometer. In the field, a MFR sun photometer directly observed the total incoming diffuse irradiance. In the laboratory, a 1000W brightness-stabilized quartz tungsten halogen lamp was used. For the first direct comparison of field and laboratory measurements, we used an artificial and inert target with high angular anisotropy. Analysis shows that the diffuse illumination in the field is leading to a higher total reflectance and less pronounced angular anisotropy.