This contribution will provide an overview of existing approaches, methods and resources for the validation of spaceborne imaging spectroscopy data products, from level 1B to level 2A. It will deal with the characterisation of the spectral, spatial and radiometric properties of the measurement and with the identification of potential issues affecting the scientific exploitation of the data. The particular characteristics of the EnMAP mission will be taken as a reference, but the aim will be to address spaceborne imaging spectroscopy missions in general.


The current version of the EnMAP validation plan includes a number of scene-based and ground-based procedures for data quality assessment. Scene-based methods include automatic algorithms based on image processing and radiative transfer principles for the evaluation of image characteristics such as spectral of smile, keystone, systematic radiometric offsets and signal-to-noise ratio. Regarding ground-based methods for product validation, it is foreseen to acquire data over selected sites worldwide which offer appropriate conditions for the evaluation of EnMAP products, including CEOS Reference and instrumented sites, sites and those from new cal/val initiatives such as Hypernets.

This contribution fits into the “Mission calibration and data validation plans and philosophy for data quality assurance and uncertainty quantification” listed among the workshop objectives. It will intend to trigger discussions on cal/val methods in general, and in particular of possible joint cal/val activities between imaging spectroscopy missions to operate around the 2020 time frame.

A major endeavour is underway for a large scale comprehensive plan for Cal/Val activities of the Italian space-imaging-spectroscopy mission PRISMA. The initiative, coordinated by the Italian Space Agency (ASI), is led by a number of Institutes of the National Research Council of Italy (CNR) that will concur to the Cal/Val scheme implementation. The plan involves 15 validation sites distributed from the Alps to the Southern Mediterranean sea and spanning from bright to dark spectral targets. The validation sites are based in agricultural areas (Grosseto-Tuscany; Arborea-Sardegna, Ferrara-Emilia Romagna and Foggia-Puglia), on coniferous forests (Rovereto-Trentino), on lakes (Trasimeno-Umbria and Garda-Lombardia), coastal waters (Lampedusa-Sicily and Venice) and snow covered surfaces (Valle d'Aosta). For each site, accurate ground measurements of radiance and reflectance on the occasions of satellite overpasses will be provided to finally compile a wide spectral database for validating PRISMA images. Relevant data on aerosol characterization for atmospheric corrections will also be provided in most of the sites. Part of the validation activity will also involve the use of an airborne platform that will be equipped with a hyperspectral sensor. This will mainly operate in the more fragmented agricultural site in Tuscany to provide spatially and spectrally resolved maps at high spatial resolution for a comparison with the satellite data. This presentation will provide information to the Cal/Val activities for PRISMA and will give details to the open resources which are made available by ASI and CNR to international collaboration, likely attracting the interest of the scientific community involved in hyperspectral observations.

For the EnMAP mission, as well as for other novel imaging spectrometer missions like DESIS, PRISMA, HISUI and others, it is of great importance to provide well-calibrated and validated data products underpinning the usefulness of spectroscopic data from space. In order to ensure high data product standards during the complete mission lifetime, operational workflows are established within the EnMAP Ground Segment.

The topics of this oral presentation are the DataQC procedures within the online processing chain, as well as the foreseen pre-launch and in-orbit calibration procedures in order to characterize and thus ensure the radiometric, spectral and geometric properties of the sensor. In addition, the foreseen data products and the underlying product models are outlined, including the newest developments to link in to the CEOS ARD CAR4L framework (CEOS Analysis Ready Data for Land). The underlying procedures within the processing chain include the generation of Data Quality Flags and rich metadata describing the quality of each acquired dataset. For the estimation of typical uncertainty budgets of the data products, the results of simulation studies studies are highlighted. Furthermore key modules of the processing chain like the PACO atmospheric correction are included in international comparison studies (i.e., ACIX and ACIX-2). To conclude, having these calibration, validation and DataQC procedures in place, the overall quality of EnMAP data products is ensured and documented.

NASA is conducting a study of satellite mission architectures suitable for observation of the Earth’s surface biology and geology (SBG). The mission will likely include at least one imaging spectrometer spanning the visible and shortwave infrared (VSWIR) wavelength region and multi- or hyperspectral imager in the thermal infrared (TIR). However, mission architecture candidates could vary in orbital configuration, number of instruments and platforms, spacecraft design and size, instrument performance and characteristics, global partnerships, etc. Science and application objectives will also drive data quality requirements, including target uncertainties of derived observations, spectral and spatial resolution, observational geometry and measurement frequency. To achieve and validate data quality requirements, the design and capacity of the calibration and validation infrastructure must be considered across candidate architectures. Such infrastructure must quantify, monitor and update radiometric, spectral, and geometric characteristics and performance of sensors both prior to launch and in orbit. Support of on-board reference systems, vicarious calibration, calibration maneuvers in space must also be considered. Deployment of identical or similar sensors to improve observation coverage and temporal frequency would require inter-calibration capabilities. Satellite observation temporal-spatial and geometric characteristics and the number and distribution of surface resources must be commensurate with vicarious calibration and the validation of dynamic process observations to target uncertainties. Comparing and selecting architectures includes assessment of the calibration and validation requirements and capabilities.

The German Aerospace Center (DLR) and the US company Teledyne Brown Engineering (TBE) have developed the hyperspectral instrument “DLR Earth Sensing Imaging Spectrometer” (DESIS), which is integrated into the “Multi-User System for Earth Sensing” (MUSES) installed on the International Space Station ISS. The instrument was launched on June 29, 2018, from Cape Canaveral with a SpaceX rocket. The instrument has been integrated into MUSES end of August 2018, which also defines the start of the commissioning phase. After the commissioning phase with functional tests, parameter tuning and performance analysis, the mission enters the operational phase. From this point on, hyperspectral data will be available to the scientific and commercial user community.

In this talk, the results of the commissioning phase will be presented including aspects of in-orbit operational performance (geometric, radiometric and spectral aspects), data quality and also first data evaluation results. In particular the procedures for and the results of the vicarious Cal/Val activities using the well-established pseudo-invariant calibration sites (PICS), the permanently equipped CEOS RadCalNet sites and in-situ Cal/Val campaigns in cooperation with CSIRO are presented. Finally, the inter-calibration with the upcoming HISUI imaging spectrometer and other EO sensors on board the ISS are briefly outlined, demonstrating the synergies between these missions.

Operational earth observation from space aims to balance users’ needs regarding spatial, spectral and temporal properties. To surpass single satellite limitations, constellation missions are set up. Multiple smaller platforms allow to increase coverage and reduce revisit time in a cost-efficient way. The challenge in hyperspectral imaging from such platforms is not just the instrument size, but lies also with data handling, radiometric and geolocation requirements.
CSIMBA (Compact Smartspectral Imager for Monitoring Bio-agricultural Areas) is a proto-flight model hyperspectral instrument being developed for an in orbit demonstration mission, with a launch target in Q4 2020. The instrument uses ‘push-frame’ imaging with a 12Mpix frame hyperspectral sensor developed within the ESA-GSTP project CHIEM. Using precise direct filter deposition, more than 150 narrow VNIR band filters are arranged in a linewise pattern over the sensor. Full hyperspectral information is built up by capturing images in rapid succession while scanning. Operationally similar to pushbroom systems, it still allows image-based co-registration using individual acquisitions.
The instrument will be fitted into a 12U CubeSat platform; The CSIMBA IOD mission will operate from 500km altitude and offer 20m GSD at 80km swath. It will use data reduction techniques to maximize its useful data return. Substantial efforts will be put into vicarious calibration methods, by including acquisitions over invariant terrestrial test sites, as well as lunar calibration. Cross-calibration with relevant instruments will build on expertise gained within the PROBA-V image quality center. The CSIMBA IOD mission will act as a precursor for a constellation of hyperspectral cubesats.