Long-range hyperspectral imaging

Fred Sigernes^1,2, Marie Bøe Henriksen ², Sivert Bakken ², Joseph Garrett ², Eirik Selnæs Sivertsen ², Roger Birkeland ², Mariusz Eivind Grøtte ² and Tor Arne Johansen ²

¹ The University Centre in Svalbard (UNIS), Norway
² Norwegian University of Science and Technology (NTNU), Trondheim, Norway

The HSI v6 is a well-tested pushbroom design based on a transmitting grating [1-3]. Other candidates were evaluated in the draft proposals [4] for this project, but due to availability, cost, and uncertainty in optical performance it was decided to use our own design. Figure 1 shows and exploded view of the instrument. All spesifications and details are listed in [2]. The instrument is presented as a do-it-yourself project using commercial off-the -shelf components.

In the following tests, the default 50 mm front objective from Edmund Optics (EO) is replaced by several candidates including a Nikon 500 mm reflex lens, a Sky-Watcher 400 mm Apochromatic refractor objective (APO), a 6 inch diameter Celestron Schmidt-Cassegrain (C6) and a 8 inch diameter Celestron Rowe - Ackermann Schmidt Astrograph (RASA 8) telescope.

In order to compare the different candidates, the HSI v6 was first used with the standard 50 mm focal length objective from Edmund Optics. Figure 2 shows a color RGB composite generated from the hyperspectral data cube, using center wavelengths 486, 557 and 630 nm representing blue, green and red color channels, respectively. Spectral resolution was set to 5 nm. An exposure time of e = 33 msec is used, which corresponds to 30 Frames Per Second (FPS). Note that the image is downsized vertically 50% and color auto balanced by Paint.net. The above is the standard procedure followed for images generated in this test.

The instrument was mounted on a rotary tablet (Syrp Genie Mini II) to sweep ΔΦ = 30° horizontally in P = 60s. The scenario is looking through an office window at UNIS towards North-East in Longyearbyen.

The Nikon 500 mm reflex telephoto lens was mounted to the HSI v6 as shown in Figure 3. The size and mass of the lens required extra mount support by an aluminum bar and a 3D printed spacer. The diameter of the lens is close to 90 mm and its about 140 mm long. The mass is 815 g. In addition, a 3D printed dovetail was made to fit the Vixen style saddle of the iOptron MiniTower II motorized telescope Goto mount.

The Goto mount was set to sweep ΔΦ = 2° in azimuth for a period (P) of about one minute. This is speed setting 8X for the motors. Angular speed is then only ω = 0.035°/s. The resulting color composite is shown in Figure 4. The distance to the target is approximately 250 meters. See target area marked with red box in Figure 2.

The F/8 aperture of the lens required a gain setting at 29.8 dB in order get a reasonable 8-bit signal count. Range is 0 - 48 dB. The image quality seems to be improved especially in the deep blue and red parts of the spectrograms. They simply look sharper than before. This behavior is expected due to the fact that the instrument is now illuminated more in the center of the lenses at F/8 compared to F/2.8. On the other hand, it was hard to focus the lens. The focus ring was very sensitive to movement / rotation.

Note that the Raspberry PI 4 is an optional mobile recording unit for the instrument. An USB3 cable may also be used directly connected to a laptop computer running Windows 11. The PI runs on the Raspbian OS with Python drivers and OpenCV. Remote control is obtained by configuring the PI as a wireless Wi-Fi hotspot and enabling the built in VNC server. Power is then supplied by a mobile phone 12Ah power bank. With this configuration the PI can operate approximately 4 hours assuming a power consumption of maximum 3A.

The use of a 2X C-mount focal length extender from Edmund Optics (#54-356) increases the focal length of the Nikon lens to 1000 mm. The extender only increases the length of the system by 11.7 mm and mass with 48 g. The down side is that the aperture is reduced by 2 full stops to F/16.

The focus was extremely challenging and the whole system was very sensitive to vibrations during scanning. See edge of solar panels in Figure 5. The extenders are known [5] to reduce image quality based on the fact that they act as a magnifying glass with increased chromatics aberrations, decrease in sharpness and contrast.

The HSI v6 connects to the APO with a C-mount to 1.25-inch adapter tube directly to the 90° mirror diagonal in the back. All parts are screw tighten by hand, which makes the assembly unstable. This should be addressed if it becomes a candidate for space. The mass is now as high as 5.1 kg and the system length with the Sun shade on becomes 520 mm.

The aperture of the APO is as low as F/5 with a lens diameter of 80 mm and a focal length of 400 mm. This is the highest light throughput candidate tested so far except from the EO 50 mm at F/2.8. Figure 7 shows the result. Again, it was hard to focus the spectrograms even with the fine adjustment focus knob option of the APO.

As with the APO, the HSI v6 is connected by a 1.25-inch tube adapter directly to the 90° diagonal mirror in the back of the C6 telescope. The configuration is unstable and should be re-enforced mechanically. The diagonal is made of plastic and should be replaced by metal. A complete disassembly of the telescope must be done do identify other parts that would not survive space.

It is suspected that the focus mechanism inside the tube is not rigid enough to resist vibrations and shocks during launch. The primary mirror slides on the center baffle tube to focus. Reports of primary mirror flops that causes image shifts, occur even using the telescope on a stationary Goto mount.

Figure 9 shows the test image obtained with the MiniTower II set to pan horizontally ΔΦ = 1.516389° in P = 3 minutes. This corresponds to speed setting 2X for the motors. Angular speed then becomes ω = 0.0084°/s.

The focus knob has a rubber cover and is very sensitive to focus adjustments. The above test image is not resized, only color balanced by Paint.net. The slow sweep period of P = 180s with 5417 recorded spectrograms resulted in close to square image pixels with no vertical binning necessary.

Expected delivery in June.

Target distance z			540 km		250 m
Front Optics	f [mm]	F/#	Δx [m]	Δy [m]	Δx [mm]	Δy [mm]
EO-50	50	2.8	551	63.3	322.7*	29.3
Sky-Watcher 80ED	400	5.0	78.5	7.91	36.3	3.66
Celestron RASA 8	400	2.0	78.5	7.91	36.3	3.66
Nikon Reflex	500	8.0	65.0	6.33	30.1	2.93
Nikon Reflex 2X	1000	16.0	38.0	3.16	17.6	1.46
Celestron C6	1500	10.0	29.0	2.11	13.4	0.98

Table 1. Spatial resolution numbers. ω = 0.035°/s for the iOptron MiniTower II. * ω = 0.5°/s for the Syrp Genie tablet. FPS = 30. The entrance slit is 0.05 x 7 mm² in size. The Sony IMX174 CMOS has 1920 x 1080 pixels. The size of the pixels is 5.86 µm square.

The above shows how to estimate the spatial resolution for different front optics assuming no movement other than rotation between target and instrument. Additional information and basic theory of operation of hyperspectral pushbroom imagers are listed in [6].

It is clear the HSI v6 may use longer focal length front optics than originally planned. The reflex mirror or catadioptric lenses are very compact and has low mass compared to traditional APO lenses with same focal length. But they are all challenging to focus by hand. In addition, the focus mechanism of the Celestron telescopes may be difficult to ruggedize for a life in orbit. Surprisingly, the old Nikon 500 mm reflex lens that was manufactured from 1983 - 2005 has equal optical performance compared to the other more modern candidates in this test. It also has the most ruggedized focus since it moves the secondary instead of the primary mirror by a fine threaded outer tube. Note that the company Tokina has released a new production line of reflex lenses in 2022.

1. M. E. Grøtte, R. Birkeland, E. Honore-Livermore, S. Bakken, J. L. Garrett, E. F. Prentice, F. Sigernes, M. Orlandic, J. T. Gravdahl, T. A. Johansen, Ocean Color Hyperspectral Remote Sensing with High Resolution and Low Latency - the HYPSO-1 CubeSat Mission, IEEE Trans. Geoscience and Remote Sensing, Vol. 60, pp. 1-19 (2022), https://doi.org/10.1109/TGRS.2021.3080175
2. M. Henriksen, E. Prentice, C. van Hazendonk, F. Sigernes, and T. Johansen, Do-it-yourself VIS/NIR pushbroom hyperspectral imager with C-mount optics, Opt. Continuum 1, 427-441 (2022), https://doi.org/10.1364/OPTCON.450693
3. S. Bakken, M. B. Henriksen, R. Birkeland, D. D. Langer, A. E. Oudijk, S. Berg, Y. Pursley, J. L Garrett, F. Gran-Jansen, E. Honore- Livermore, M. E. Grøtte, B. A. Kristiansen, M. Orlandic, P. Gader, A. J. Sørensen, F. Sigernes, G. Johnsen and T. A. Johansen, HYPSO-1 CubeSat: First Images and In-Orbit Characterization, Remote sensing, 15(3), 755 (2023), https://www.mdpi.com/2072-4292/15/3/755
4. Fred Sigernes, Marie Bøe Henriksen, Sivert Bakken, Joseph Garrett, Eirik Selnæs Sivertsen, Roger Birkeland, Mariusz Eivind Grøtte and Tor Arne Johansen, Proposal drafts: [I] and [II] HYPSO-3 Hyper Spectral prototypes, Norwegian Space Agency, January 25, 2023.
5. Dan Carr, Shutter Muse, January 5, 2020, Tutorial: The Ultimate Guide To Extenders Or Teleconverters, https://shuttermuse.com/ultimate-guide-to-extenders-teleconverters/
6. Hyper spectral Imaging, [pdf], F. Sigernes, AGF-331: Remote Sensing and Spectroscopy - lecture notes, University Centre in Svalbard (UNIS), 2000 - 2007.