1.1

The EnMAP Instrument

The EnMAP HSI is a state-of-the-art satellite-based hyperspectral Instrument for Earth observation in push broom configuration. The full spectral range is measured with two separate spectrometers, VNIR and SWIR, which share the same telescope, but have different line of sight. In both spectrometers the EnMAP swath width of 30 km is sampled in spatial direction with 30 m ground sampling distance. The main performance parameters are summarized in Table 1.

Figure 1.

EnMAP Field Of View (FOV) and overpass concept.

Table 1.

Key specifications for the EnMAP instrument.

Figure 2 shows a cross-section view of the HSI. The instrument features a nearly diffraction-limited three-mirror anastigmat telescope assembly (TA) with an across-track Field Of View (FOV) of 2.63°. The TA has a 180 mm entrance aperture and focuses the light from the earth’s surface onto a field splitter slit assembly (FSSA). The FSSA includes two separate micro slits for in-field separation of the light and a micro-mirror to redirect the SWIR field into the SWIR spectrometer [2]. Both, VNIR and SWIR spectrometer optics employ curved prisms in dual-pass configuration as dispersive elements. Light from the spectrometer is finally focused on VNIR and SWIR focal planes arrays that acquire images at a frame rate of 230 Hz with 14-bit resolution.

Figure 2.

Cross-section of the EnMAP HSI main optical units

All optical elements are mounted to a monolithic 3-dimensional Aluminum structure. The Aluminum structure together with the Spectrometers and FSSA forms the Instrument Spectrometer Unit (ISU). To avoid a loss of optical performance due to the thermal expansion of the Aluminum structure, the whole HSI is temperature controlled by a Thermal Control SubSystem (TCSS) to 21°C.

For fulfilling the stringent requirements on overall pointing stability, the star tracker sensor assembly is directly attached to the optical unit. In addition, an on-board calibration system will be used for recurring in-flight radiometric and spectral calibration. It consists of a solar calibration assembly containing a diffuser which reflects sunlight into the Telescope entrance aperture as well as a calibration light source (not shown in Figure 2) that can be fed into the beam path in front of the Spectrometer entrance slits [1]. The calibration light source is used to monitor the calibration validity at small time intervals, whereas the solar calibration assembly will be used less frequently, but as the primary radiometric calibration device during space operation.

1.2

General setup aspects

Aligning and calibrating a state-of-the-art hyperspectral instrument as the EnMAP HSI requires to establish measurement setups that outperform the test object in all relevant performance aspects to achieve the required measurement accuracies. At the same time technical as well as economic considerations yield to develop measurement equipment that can support multiple use cases throughout the Alignment integration and Test (AIT) Process of EnMAP and future projects. In this spirit, the following main aspects were considered.

2.1

Spectrometer alignment

The objective of the spectrometer alignment was the optimization of the Wave Front Error (WFE) after interface generation [4] and the placement of the optical elements (OE). The alignment process accounted for the discrepancy between desired and actual positions of the OEs due to the manufacture and placement tolerances. The overall flow of the process:

The alignment was performed for the VNIR and SWIR spectrometers separately. The WFE was measured at several field points and several discrete wavelengths in double-pass configuration with the setup depicted in Figure 3. A discrete wavelength light source equipped with 5 low coherence laser sources was coupled via fiber to the Spectrometer Spot Illumination and Detection (SpecSID) GSE. The SpecSID allowed to illuminate the entrance slit of the spectrometer at different positions with a well-defined spot illumination. From the spectrometer focal plane, the light was reflected back using a SpecSID Wavefront Reference (SIDWaRe) GSE to realize a double-pass WFE measurement. The double-pass WFE was measured using a Wavefront sensor inside SpecSID.

Figure 3.

Setup for Spectrometer alignment

2.2

Telescope assembly (TA) to Instrument Spectrometer Unit (ISU) alignment

At first, the EnMAP telescope assembly was integrated, aligned and characterized by OHB on sub-assembly level with residual (single pass) wave front error of 28 nm rms [5]. The complete assembly was later aligned to the ISU forming the HSI.

Figure 4 shows the setup for this TA to ISU alignment. The TA entrance pupil was illuminated with a collimated light beam of 200 mm diameter coming from the Full Aperture Illuminator (FAI) [3]. The light was focused by the TA onto the FSSA and was reflected back by micro-mirrors on the FSSA [2]. There were six mirrors corresponding to extremal field points.

Figure 4.

Setup for the Telescope to Instrument Spectrometer Unit alignment

The FAI was developed, manufactured, integrated, aligned and tested upon OHB Specifications by Bertin Technologies. It was composed of a two-mirrors unobscured collimator, a scene generator and a source assembly. The source assembly included an internal light source and allowed to analyze the reflected light for collimation. Using a scanning system manufactured by Symetrie, the collimator was displaced angularly to address the six mirrors on the FSSA. The TA was mounted on a Telescope to ISU Integration GSE (TIIG), which moved the TA in front of the Spectrometer until the best focus position of the TA regarding all six micro mirrors was derived with an accuracy of ca. 1 μm in focus direction. In this optimized position, the necessary shim thicknesses were determined. A displacement probe GSE helped to reference the TA position during shim integration.

2.3

Camera alignment

For the optical alignment of the cameras, a monochromatic spot illumination of the spectrometer slit was used, similar to the Spectrometer and the TA to ISU alignment. The light was detected directly by the flight models of the cameras, using the supporting Instrument Central Processing Unit (ICPU) and GSEOS.

For both channels, VNIR and SWIR, the same activities have been performed. First, an approximate lateral position of the camera has been found. After that focal position has been adjusted, optimizing the Modulation Transfer Function values (MTF) obtained from a series of measurements at different spectral and lateral fields. Finally, lateral position was updated to optimize usage of the detector active area and to minimize keystone.

The availability of both, SpecSID and FAI OGSE, made it possible to keep the whole alignment process flexible and allow for parallel work on the ISU and TA. According to schedule demands, the VNIR Camera was aligned and characterized using the SpecSID without the TA whereas the SWIR Camera alignment was done using the FAI after the TA final integration.

3.1

Radiometric and Polarization setup

Although a radiometric characterization was already performed on detector level, it was repeated on instrument level during the CnC campaign, to achieve true end-to-end calibration data for the instrument.

The setup is depicted in Figure 5. The main components are:

Figure 5.

Setup for radiometric calibration, polarization sensitivity characterization and SNR measurements.

The FIS had a diameter of 1.6 m and delivered homogenous broadband illumination. It was capable of illumination at 14 different power levels, which was of importance especially for HSI linearity analysis and SNR measurements. The FIS aperture of 200 mm x 300 mm was marginally bigger than the instrument FOV. For an exact absolute radiometric calibration of every pixel, the HSI FOV had to be covered without clipping. For this reason, the FIS aperture was aligned with respect to the HSI FOV with accuracy of 2 mm. The radiance homogeneity within the FIS aperture was characterized in advance to be better than 0.3% peak to valley [7].

In Figure 6, the full radiometric calibration chain is depicted: The Transfer Spectrometer (TS) was used as a calibrated radiometric reference for all measurements. The TS itself was calibrated before and after the campaign using the RAdiometric STAndard (RASTA) at DLR-IMF, which is itself calibrated to national standards at the National Metrology Institute of Germany (PTB). With this calibration chain, a final radiometric measurement accuracy of <1% standard deviation was achieved.

Figure 6.

Tracability chain for the absolute radiometric calibration of EnMAP [7].

To characterize the polarization sensitivity, a rotatable full aperture polarizer (FAP) with 280 mm clear aperture was placed and aligned between FIS and EnMAP HSI. The ultra-broadband polarizer guaranteed a degree of polarization of better than 1000 within the whole EnMAP range of 420 nm to 2450 nm.

The stability of FIS radiation, especially during the longer SNR and Polarization measurements, was monitored using two high sensitive Si and InGaAs radiometers, for the VNIR and SWIR range, respectively. Figure 7 shows the data of both radiometers during a time period of one hour at FIS level number 5. The data shows excellent stability of <0.03% standard deviation in both wavelength ranges. With the radiometers and the FAP, also the inherent degree of polarization of the FIS was validated to be smaller than 0.8%.

Figure 7.

Radiance stability of FIS measured in the VNIR and SWIR range over 1 hour.

3.2

Geometrical and spectral setup

Figure 8 shows the setup for the geometrical and spectral measurements. During the geometrical and spectral measurements, the HSI was illuminated using the FAI combined with the Wide Range Adjustable Light source (WiRAL), which was coupled to the FAI via optical fibres. To homogenize the HSI pupil illumination, a Fiber based Mixing System (FMS) was introduced in the optical fibre connection between WiRAL and FAI.

Figure 8.

Setup for geometrical and spectral measurements

The core of the WiRAL is a commercially available Grating Monochromator (iHR320) by HORIBA Scientific, fed by several light sources, which was extensively characterized and optimized by OHB. It allowed adjustable monochromatic illumination in the complete EnMAP range from 420 nm to 2450 nm and even beyond, with a bandwidth of 0.5 nm to 1 nm, which was smaller than 10% of the instrument spectral sampling distance.

WiRAL was calibrated traceable to national standards using a PTB-calibrated Echelle spectrometer provided by DLR-IMF. It had a maximum wavelength uncertainty of 0.005 nm at 1000 nm and 0.0125 nm at 2500 nm [8], which was far below the CnC requirement of 0.1 nm in the range of 420 -1000 nm and 0.2 nm in the range of 1000 -2450 nm. Additionally, a dedicated recurring calibration plan employing Gas-Cell line sources ensured the validity of the calibration throughout the complete EnMAP on-ground calibration campaign.

By scanning the WiRAL wavelength and synchronous image takes of the instrument detectors, the spectral response function of all spectral pixels of several across-track pixel rows were obtained. Since the FAI illuminated the HSI with collimated light, a sufficient intensity level on the detector was easily achieved, but only a few pixels in across-track direction could be illuminated at the same time. Hence, after each wavelength scan, the FAI had to be moved angularly with its scanning system to address the next row of pixels, resulting in a very long measurement series. However, the implemented automatization capabilities described in section 1.2 enabled scripted measurement sequences controlled by the GSEOS software which run independently overnight without operator control. The geometrical measurements, which cannot be automatized in such a high degree, were performed in parallel during the daytime. This enabled to keep the tight CnC campaign schedule without compromises on the desired amount of measurements and without the need of shift work.

For the geometrical measurements, knife edge structures situated in the scene generator of FAI were imaged onto the instrument detectors and stepped across the illuminated field in sub pixel steps. Detector images were synchronously taken using the instrument detectors, to determine MTF and instantaneous FOV and to correlate the pixel centre with the knife edge position [3].

To correlate the HSI LoS with the FAI knife-edge, a special method has been used (a detailed description of the procedure can be found in [3]). The method implied a Full Aperture Mirror (FAM) aligned in front of FAI and a High Dynamic AutoCollimator (HDAC) aligned on the HSI reference cube. The FAM and the HDAC linked the optical axis of the FAI collimator with the HSI optical axis. The HDAC is a customized electronic autocollimator developed by Möller-Wedel-Optical GmbH upon OHB specifications with an accuracy of <1.2 μrad in each axis, calibrated traceable to national standards, and a resolution of <1μrad.

A special FAI monitoring system traced the angular stability of the FAI collimator system with respect to the FAI granite base. Long-term measurements of the granite monitoring against FAI collimator, FAI collimator against FAM, FAM against HDAC and HDAC against HSI proved an overall setup stability of better than 1 μrad over several hours.

FAI collimator WFE measurements were done using the FAM as a flat reference surface. A regular check of the FAI showed a stable WFE of <60 nm rms over its 3 years of usage at OHB.

For the Knife-edge measurements (as well as for spectral characterization) a very homogeneous light field, both in real (focal plane-) space as well as in angular space, is desired. Figure 9 shows the illumination homogeneity in the focal plane of FAI measured with an Avalanche Photo Diode (APD). The two-dimensional spot in the X-Y plane was scanned along X for different Y positions. The measurement showed a perfect homogeneity of better 2% PV due to the used fiber illumination.

Figure 9.

Illumination homogeneity in FAI focal plane, scanned with Avalance Photo Diode (APD).

To achieve high spectral calibration accuracy of HSI, the WiRAL had to deliver stable optical power during a scan range of at least 3 SSD (ca. 5 min). Figure 10 shows a stability measurement and analysis of WiRAL output power at 970nm over 3 hours. In Figure 10a, the evolution of the measured power E is shown directly after switch on of the lamp. From the measured power values within a time interval of 5 min the trend and the normalized random noise (NRMSD) is calculated using the equations (1) to (3).

Figure 10.

WiRAL output power stability at 970nm. a) relative optical power measured with a Si radiometer. b) Determined trend and NRMSD evolution over time.

Figure 10b shows the change in trend and NRMSD over time. After 0.6 h (36 min) heat up time, the required stability of 0.5% trend and 0.2% NRMSD was achieved and dropped further to a stability of <0.2% trend and 0.05% NRMSD for long term usage.

3.3

Straylight setup

The setup for the straylight measurements was very similar to the setup for the spectral and geometrical measurements and is depicted in Figure 11. The FAI was used to illuminate the HSI together with WiRAL and the Stray Light Test Source (SLTS) from DLR-IMF as light sources. All GSE between FAI and instrument (HDAC, FAM) were removed to reduce possible stray light sources close to the EnMAP FOV. Additionally, a CAPE® tent with opaque fabric was set up around the setup to reduce parasitic light e.g. from small indicator LEDs of other devices in the room (see Figure 12).

Figure 11.

Stray light setup

Figure 12.

Picture of ISO5 cleanroom setup.

As a very versatile OGSE, the FAI was designed right from the beginning to be also suitable for stray light measurements. The FAI scene generator was designed to allow special illumination scenes on the FSSA for stray light measurements, like 2 x 2 pixel illumination for the in-band stray light measurements or single pixel row illumination for spatial in-band stray light measurements.

The SLTS was designed to measure Point Spread Functions (PSF) at specific wavelengths with a very High Dynamic Range (HDR) [9]. These HDR PSFs can be used to correct the EnMAP Data for in-band stray light. Broadband illumination was possible as well, which was used for the out-of-field stray light characterization. Also the SLTS was controlled by the GSEOS software, which enabled automated measurements of HDR PSFs at numerous positions on the EnMAP detectors. On the VNIR detector, for example, HDR PSFs with a dynamic range of 10^-9 were measured at 132 positions on the detector with a pure measurement time of about 16 hours in total.

For the spatial and spectral in-band ghost characterization, the WiRAL was used as a light source, because it delivered spectral narrow band illumination adjustable over the whole EnMAP range, as well as spectral broadband illumination optimized respectively for the two VNIR and SWIR spectrometer wavelength ranges.