TSC-1 Offner Spectrometer Prototype Characterization

Abstract

The Thai Space Consortium (TSC) has undertaken the development of an Offner spectrometer prototype for the TSC-1 satellite mission, aiming to enhance Earth observation capabilities. Through systematic parameter selection and radiometric analyses, optimal performance of the hyperspectral imager within established specifications was achieved in the previous study. The design phase involved selecting a two-mirror off-axis telescope coupled with the Offner spectrometer for its diffraction-limited performance. Rigorous testing validated the prototype’s alignment with simulated performance, affirming its ability to meet demanding Earth observation requirements. The experimental results demonstrate that the Offner spectrometer prototype has been successfully developed. The spatial resolution ranges between 21.0 and 24.1 µm, and the spectral resolution ranges between 7.3 and 8.7 nm, with no significant distortion. Furthermore, the signal-to-noise ratio at 550 nm is 100. This achievement positions the TSC at the forefront of innovative Earth observation instrumentation in Thailand, with implications for future space missions requiring precise and efficient hyperspectral imaging.

1. Introduction

The Thai Space Consortium (TSC) stands as a pioneering force in the realm of space exploration and satellite technology in Thailand. Established with a visionary mandate, the TSC has been at the forefront of driving innovation, fostering collaboration, and advancing the capabilities of the nation in the domain of space science and technology. Founded on the principles of cooperation and knowledge-sharing, the TSC brings together a collective of experts, researchers, and institutions dedicated to pushing the boundaries of space research. With a strategic focus on developing and launching Earth observation satellites, the consortium envisions a future where Thailand plays a vital role in the global space community.

The TSC serves as a platform for interdisciplinary collaboration, where experts from diverse fields such as aerospace engineering, optics, electronics, and data science converge to create state-of-the-art space technologies. By pooling resources, knowledge, and expertise, the consortium aims to address the unique challenges of space exploration and contribute meaningfully to scientific discoveries and advancements. One of the flagship projects under the TSC umbrella is the development of the TSC-1 Earth observation satellite [1]. This ambitious initiative reflects the consortium’s commitment to harnessing the power of space-based technologies for applications ranging from environmental monitoring to resource management. The TSC-1 project symbolizes Thailand’s entry into advanced space technology and Earth observation, marking the start of a new era of scientific and technological progress that fills the nation with pride.

In addition to its focus on satellite technology, the TSC is actively involved in promoting education and awareness in space-related disciplines. Through training programs, workshops, and outreach activities, the consortium is nurturing the next generation of space scientists and engineers, ensuring a sustainable and knowledgeable workforce for Thailand’s space endeavors.

The TSC-1 payload represents an advancement in hyperspectral imaging technology, featuring a compact design optimized for small satellite platforms. This innovation supports more frequent and flexible Earth monitoring at a lower cost, bridging gaps between larger satellite missions. Its improved signal-to-noise ratio (SNR) ensures higher data quality and reliability, minimizing the need for extensive post-processing corrections. The TSC-1 addresses critical scientific questions, such as improving the detection and monitoring of environmental changes, supporting sustainable agriculture practices through detailed insights into crop health and soil properties, and aiding urban planning by monitoring urban heat islands, vegetation cover, and infrastructure development. Moreover, the TSC-1 opens new research avenues by enabling detailed vegetation mapping crucial for biodiversity studies and conservation efforts, enhancing coastal water quality monitoring through detecting chlorophyll concentration, suspended sediments, and pollutants, and supporting sustainable mineral exploration by providing detailed information on mineral composition and distribution. These capabilities demonstrate the TSC-1’s substantial contributions to advancing Earth observation science, showcasing its significance beyond mere technical specifications.

In the design phase, we conducted a bibliographic review of 30 airborne and spaceborne hyperspectral imager designs [2], which led to the selection of the Offner configuration due to its remarkable optical performance with no signification distortions and its simple design.

To provide a robust scientific foundation for the TSC-1 hyperspectral imager, it is important to contextualize its advancements within the broader field of hyperspectral imaging. Recent developments have seen significant improvements in both spaceborne and airborne hyperspectral imagers, each contributing unique capabilities to Earth observation. Among these, notable systems include AVIRIS-NG [3], Hyperion [4], PRISMA [5], Environmental Mapping and Analysis Program (ENMAP) [6], DLR Earth Sensing Imaging Spectrometer (DESIS) [7], and Advanced Hyperspectral Imager (AHSI) [8], which offer valuable benchmarks for assessing the performance and innovations of the TSC-1 hyperspectral imager.

AVIRIS-NG excels in providing high-resolution hyperspectral data, which are invaluable for a wide range of applications such as assessing vegetation health, detecting environmental pollutants, and mapping mineral resources [3]. Its ability to capture detailed spectral information across a wide range of wavelengths allows for precise identification and analysis of surface materials, making it a critical tool for researchers and environmentalists. However, as an airborne platform, AVIRIS-NG is limited to specific missions and temporal coverage, and cannot offer the continuous, long-term monitoring that satellite-based sensors provide. This limitation restricts its suitability for comprehensive, large-scale environmental monitoring projects. In contrast, the TSC-1 hyperspectral imager, being satellite-based, delivers continuous and extensive coverage necessary for sustained and wide-ranging monitoring efforts.

Hyperion on the EO-1 Satellite stands out with its extensive spectral range covering the visible to near-infrared (VNIR) and shortwave infrared (SWIR) regions [4], making it suitable for diverse applications such as vegetation analysis, mineral mapping, and environmental monitoring. However, its lower spatial resolution and reliance on older technology can limit its effectiveness in applications requiring finer detail and more modern data acquisition methods. In comparison, the TSC-1 hyperspectral imager offers more contemporary technology with improved spatial resolution, making it a stronger candidate for applications needing greater detail and accuracy.

PRISMA offers a significant advantage with its high spectral resolution and broad spectral range, enabling precise and detailed analysis for a wide range of scientific and industrial purposes, including agriculture, forestry, and urban planning [5]. Its advanced data processing capabilities facilitate complex data interpretation and application. Despite these benefits, PRISMA’s larger platform and higher associated costs may present challenges for some users, particularly those with budget constraints or requiring more compact and economical solutions. The TSC-1, while advanced, boasts a more compact design and potentially lower costs, making it a more accessible option for a broader range of users.

ENMAP is designed to provide high-resolution hyperspectral data for a wide range of applications [6], such as monitoring water quality, assessing land degradation, and tracking ecosystem changes. Its high spectral resolution and advanced data processing capabilities make it a powerful tool for detailed environmental analysis. However, like PRISMA, ENMAP’s larger platform and associated costs might pose challenges for some users. In comparison, the TSC-1 offers similar capabilities in a more compact and potentially cost-effective package, making it an appealing alternative for those requiring high-resolution hyperspectral data without the larger infrastructure.

DESIS on the International Space Station (ISS) and the TSC-1 spectrometer offer distinct advantages through their respective operational platforms and technical specifications. DESIS, with a spectral range of 400 nm to 1000 nm and a spectral resolution of 2.55 nm [7], is optimized for detailed hyperspectral data acquisition in various applications. Its installation on the ISS provides significant observational flexibility, allowing it to take advantage of frequent revisit times and diverse viewing angles. This capability is particularly beneficial for capturing dynamic and short-term changes on the Earth’s surface, making DESIS a valuable tool for applications requiring high temporal resolution. On the other hand, the TSC-1 hyperspectral imager, while providing 10 nm resolution hyperspectral data over the same spectral range as DESIS, is designed for continuous and extensive global monitoring from its dedicated satellite platform. TSC-1 provides long-term, consistent data critical for large-scale environmental and climate studies, ensuring sustained observational capabilities essential for monitoring gradual changes and long-term trends.

On the other hand, AHSI is configured in an improved three-concentric mirror Offner system [8], which is more complex and more sensitive to misalignment than the TSC-1 spectrometer. The AHSI on Gaofen-5 provides a broad spectral range from 400 nm to 2500 nm with high spectral resolution (5 nm in the visible to near-infrared range and 10 nm in the shortwave infrared range [8]), enabling it to perform detailed analyses across a wide range of wavelengths. In contrast, the TSC 1 hyperspectral imager covers a visible spectral band with 10 nm resolution with a simple configuration to enable a more robust system.

Overall, while AVIRIS-NG, Hyperion, PRISMA, ENMAP, DESIS, and AHSI each have their strengths and limitations, the TSC-1 hyperspectral imager combines modern technology, improved spatial resolution, cost efficiency, and continuous coverage, making it a versatile and effective solution across various applications and user requirements.

1.1. Background Context

The design and performance prediction of the TSC-1 hyperspectral imager were guided by rigorous theoretical models and detailed simulations. We employed Zemax OpticStudio for the optical design and performance estimation of the spectrometer. Zemax OpticStudio 24.1 is a powerful optical design software widely used in the industry, enabling precise modeling of complex optical systems. The design process involved several key steps, including the selection of appropriate optical components, optimization of optical paths, and thorough analysis of the optical performance metrics. The optical principles underlying the design of the TSC-1 hyperspectral imager are based on the Offner relay configuration, which is known for its high spectral resolution and low aberration properties. This configuration was chosen due to its ability to provide a compact and efficient optical path, which is critical for spaceborne applications where size, weight, and power consumption are tightly constrained. Several assumptions and approximations were made to streamline the simulations during the modeling process. For instance, we initially assumed ideal optical surfaces to evaluate the baseline performance of the design. Realistic surface imperfections, manufacturing tolerances, and alignment errors were introduced in later stages to assess their impact on optical performance. These factors were incorporated into the tolerance analysis to ensure that the final design meets stringent performance requirements even under non-ideal conditions. Furthermore, the simulations accounted for the thermal environment of space, which can induce significant thermal stresses and deformations in the optical components. Thermal analysis was performed using finite element modeling (FEM) techniques to predict the behavior of the optical system under varying thermal conditions. This analysis helped optimize the mechanical design to mitigate thermal effects and maintain optical performance. In addition to optical and thermal simulations, stray light analysis was conducted to evaluate and minimize the impact of unwanted light on the hyperspectral images. A comprehensive baffle design was implemented based on the stray light simulation results to protect the optical system from external light sources. Overall, the combination of advanced optical design software, detailed theoretical modeling, and extensive simulations provided a robust framework for the development of the TSC-1 hyperspectral imager. These efforts ensured that the spectrometer would deliver high-quality hyperspectral data for Earth observation, addressing specific scientific questions and opening new avenues for research in remote sensing.

1.2. Offner Spectrometer Development

The optical payload of the TSC-1 Earth observation satellite marks Thailand’s inaugural foray into optical space systems for small satellites. In ensuring the payload’s feasibility, we made deliberate technical decisions. Initially, we prioritized using spherical or aspherical optical surfaces, over freeform optics in our first payload. Additionally, we confined the spectral band to the visible domain.

The development of an Offner spectrometer endeavor represents a crucial component within the broader TSC-1 Earth observation satellite initiative, showcasing the consortium’s dedication to advancing optical systems for space applications. The Offner spectrometer, a sophisticated optical instrument with unique advantages in hyperspectral imaging, is being meticulously crafted to meet the stringent demands of Earth observation from space. The Offner configuration, characterized by its compact design and reduced optical aberrations, is particularly well-suited for the challenges posed by satellite-based hyperspectral imaging.

To achieve optimal performance, the TSC has implemented a multi-phase approach to developing the Offner spectrometer. This includes a comprehensive survey of existing hyperspectral imaging technologies to inform parameter selection [2], radiometric analyses to estimate SNR, and a careful selection of aperture numbers to ensure the instrument’s efficiency across specified spectral intervals [1].

We aim to develop the prototype of the Offner spectrometer to validate its theoretical design and compare it with the simulation results. The Offner spectrometer, known for its distortion-free imaging capabilities and its compacity [9], holds significant promise for Earth observation and other applications. This study delves into the design and construction of the prototype, optical component material selection, and optical system integration for optimal performance. The laboratory experiments were conducted to assess the prototype’s spatial and spectral resolution and distortions. Challenges encountered during development are addressed, showcasing the instrument’s ability to capture spectroscopic data. This study serves to bridge the gap between theoretical simulation and practical realization, demonstrating the potential of the Offner spectrometer for future applications. It is important to mention that the current paper focuses exclusively on the development and technical aspects of the Offner spectrometer prototype. The processing and analysis of the acquired hyperspectral images, including any subsequent methodologies will be reserved for a subsequent publication dedicated specifically to that topic.

2. Prototype Mechanical Design Strategy

2.1. Material Choice

We have a strategy to use a single material for developing this instrument, not only for the BreadBoard Model (BBM) phase but throughout the Flight Model (FM) phase. A single-material approach can mitigate thermal expansion mismatches that often occur when multiple materials with different coefficients of thermal expansion are used in the same system. Thermal stability is critical for space instruments, as they are exposed to extreme temperature variations between sunlit and shaded regions [10]. Using a single material minimizes the effects of temperature-induced dimensional changes, thereby maintaining the optical system’s precise alignment and preserving its performance in challenging space environments. Therefore, we considered using aluminum for both the mechanical structure and optical components. One challenge of using aluminum optics is achieving a high-quality optical surface finish on aluminum, which can be difficult due to its softness and tendency to form a rough surface during the polishing process. However, using diamond turning techniques can produce high-quality optical surfaces on aluminum. Post-polishing treatments, e.g., electroless nickel plating or a protective coating, can also improve the surface finish, oxidation, and durability [11].

2.2. Optical Components

The main optical components, which are the concave mirror (SPECTRO-M1) and the grating, were machined by the Center for Advanced Instrumentation (CfAI), Durham University using a diamond turning tool on an aluminum block. The measured parameters of these optical surfaces are presented in

Table 1. SPECTRO-M1 surface quality.

Parameter	Specification	Measurement
Mechanical diameter	101.6 mm +0.000/−0.025	101.594 mm
Clear aperture	90 mm ± 0.1 mm	90 mm
Radius of curvature	102.181 ± 0.1%	102.200 mm
Surface irregularity	<150 nm (PTV)	81 nm (PTV)
Roughness	3–4 nm	3–4 nm RMS
Scratch Dig pre-silver coating	ISO 10110-7:2017 [12]	compliance
Scratch Dig post-silver coating	ISO 10110-7:2017 60-40 [12]	Visual inspection
Substrate material	Aluminum 6061 T6 RSA [13]

PTV: peak to valley, RMS: root mean square.

for the SPECTR-M1 and

Table 2. The convex grating surface quality.

Parameter	Specification	Measurement
Thickness	22.54 ± 0.02 mm	22.560 mm
Clear diameter	20 ± 0.1 mm	20 mm
Radius of curvature	51.481 mm ± 0.1%	51.485 mm
Grating density	41 lines/mm	41 lines/mm
Blaze angle	optimized for $\lambda$ = 500 nm at diffraction order +1	0.724° ± 0.05°
Surface irregularity	<633 nm (PTV)	586 nm PTV
Roughness	<3 nm RMS	3 nm RMS
Scratch Dig	ISO 10110-7:2017 60-40 [12]	No surface defects
Substrate material	AlSi40 RSA443 + NiP

PTV: peak to valley, RMS: root mean square.

for the grating. Although the grating surface irregularity is approximately 600 nm, it will be demonstrated in the subsequent section that its performance meets the spectral requirements. The entrance slit mechanical diameter is a 1-inch circular shape to fit with the commercial mount. The image of the slit covers 2 pixel rows, 22 µm in the transverse direction of the slit. The slit length is 11 mm covering 30 km FOV when it is equipped with the TSC-1 front telescope.

2.3. Mechanics

In order to provide a stable and precise mounting platform for specific commercial kinematic mounts, we designed a dedicated mechanical base (spectro-base). This mechanical base serves as a mechanical reference and provides a solid foundation [5] for the secure attachment and alignment of the commercial kinematic mounts. Commercial mounts were selected for prototype development due to their cost-effectiveness, immediate availability, proven reliability, and adherence to industry standards. They also offer ease of use with comprehensive documentation, technical support, and warranties. These mounts enable quicker integration and testing, facilitate scalability, and allow the development team to focus on core project aspects, ensuring a more efficient and successful development process. However, as the project progresses, custom mounts will be developed in the future to meet the specific structural requirements and ensure compliance with space standards. The design of the mechanical base considers many factors such as material selection, structural integrity, and dimensional accuracy. By carefully designing and manufacturing the spectro-base, we ensure that it provides the necessary rigidity and stability to support the kinematic mount, minimizing any potential mechanical vibrations or movements that could adversely affect the performance of the optical system. The spectro-base is made of an aluminum block with less than 10 µm precision.

Three key components, the entrance slit, the SPECTRO-M1 and the grating, are attached to this mechanical base. A right-angle prism mirror is an additional component employed only for this prototype, as presented in Figure 1. Because of the limitedavailable space, a folding mirror is used to accommodate the detector. However, in the following development phases, our developed detector will take place with a more compact form factor and be compatible with the available space. Therefore, a folding mirror will no longer be employed. This makes the system more robust and stable.

2.4. Detector Choice

According to our bibliographic study [1], many new-generation hyperspectral imaging systems use the AMS CMOSIS CMV series detector [14,15]. This detector series provides high-image quality and low noise. It is capable of a high frame rate and low power consumption. Moreover, it has good temperature performance, allowing it to operate effectively in various environments, including extreme temperatures. In our case, we use an NIR-enhanced monochrome CMV-4000 model that works across the 400 nm to 1000 nm spectral range.

3. Mechanical and Optical Alignments

The Offner spectrometer is an off-axis system. Therefore, it is difficult to align [16,17]. In case of slight misalignment, optical aberrations are easily introduced. To be able to align the setup, understanding the light path is essential. Wrong adjustments might lead to a more difficult alignment method. In this setup, the light from the light preparing unit (LPU) passes through the slit and enters the spectrometer unit. The light is collimated by the SPECTRO-M1 and reflected toward the grating. The grating diffracts the light into a spectrum and reflects it toward the SPECTRO-M1. Finally, the spectrum is focused by the SPECTRO-M1 onto the detector at the image plane.

There are five main components, which are a linear slit, SPECTRO-M1, the grating, a right-angle prism mirror, and a detector. During assembly, these components are placed in their mounts before putting them on the spectro-base. The spectro-base is designed in cooperation with the mechanical design of the commercial mounts we planned to use. Therefore, when these components are installed onto the spectro-base, they are roughly in their theoretical position. With a minor adjustment, they will be in their place without spending hours aligning. It is essential to mention that the grating is fixed and is a reference point.

First, the slit orientation is adjusted to be vertically straight by looking at the image provided by the sensor. After that, the slit tip/tilt is adjusted so that the long edges of the 0th-order slit image are in focus. Then, SPECTRO-M1 tip/tilt is adjusted so that the 0th-order slit image is located perfectly at the center of the sensor as designed. The slit located at the center of the sensor measurement is presented in Figure 2.

4. Light Preparing Unit Setup

To characterize the spectrometer, the LPU plays an important role in preparing the identical light characteristics as provided by the TSC-1 front telescope. The LPU ensures that the spectrometer receives a well-prepared input signal by carefully controlling factors like the light’s intensity, direction, wavelength range, and numerical aperture of the beam. This step enhances the overall performance and reliability of the spectrometer, enabling it to capture precise data and deliver meaningful results for various applications. Therefore, this is to ensure that the LPU does not degrade the optical performance of the spectrometer.

4.1. Point-Object Light Preparing Unit

To provide a point source entering the spectrometer slit for characterization, we employed three main components, which are two off-axis parabolic (OAP) mirrors [18] and one iris diaphragm. They are configured as shown in Figure 3. Instead of a refractive system, mirrors are used to avoid chromatic aberration. This approach ensures that various wavelengths are focused at the same position at the entrance slit, allowing better spectral resolution and reliable analysis of the collected data. The focal length of the off-axis parabolic mirror is chosen according to the aperture number of the TSC-1 front telescope. The focal length of the OAPs is equal to 6 inches. In corroborating with a 2-inch diaphragm, the beam aperture is emulated, as provided by an F/3 system. The light is fed by a 25 µm core optical fiber as a part of the LPU. The image provided by this optical system is at the entrance slit level. It is important to mention that this LPU cannot provide an ideal point but a given size corresponding to the optical fiber core. The optical fiber core size is well selected to match the smallest resolvable object by the Offner spectrometer prototype. After the point-object LPU was set up, a detector was placed at the focal point of the LPU to ensure the fiber core was well imaged with no significant aberration due to the LPU. Careful calibration was performed by adjusting the alignment of two OAPs to achieve a sharp and clear image of the fiber core, thereby maintaining the integrity of the light source for subsequent measurements.

4.2. Long-Slit Light Preparing Unit

On the other hand, a full slit illumination is needed to characterize the full FOV of the spectrometer. The Edmund lens set, with an F/3 and 50 mm focal length, is used as a front telescope. The object is a white flat screen. This screen is uniformly illuminated by a known source, as shown in Figure 4. To simplify the operation, the lens is configured to achieve 1:1 magnification. Both the object and image distances are equal, set at 100 mm. In the same way as setting up LUP for the point-object LPU, a detector was placed at the image plane of the Edmund lens to verify that the LPU did not introduce any significant aberrations and distortions. Moreover, it is important to verify that the flat screen image on the detector is evenly illuminated. This setup helps characterize and evaluate various aspects of the spectrometer’s FOV, such as uniformity of response, spectral resolution, and any potential variations or distortions in the measured data across the FOV. These characterization data are important for understanding the overall performance and reliability of the instrument, ensuring that it provides accurate and consistent measurements throughout its entire FOV.

5. Measurement Method

Spatial Resolution, Spectral Resolution, and Distortion Measurements

First, we start measuring by using the point-object fore optics setup equipped with the spectrometer. The spectro-base is attached to a lab jack, a vertical adjustable stage. Moving this stage up and down allows us to gather samples of a point image across the FOV, as shown in Figure 5. This method provides useful results to evaluate many essential parameters, which are spatial resolution, spectral resolution, smile distortion, and keystone distortion. The spatial and spectral resolutions can be evaluated by measuring the Full Width at Half Maximum (FWHM) of point source images across spatial and spectral directions, respectively. The smile and keystone distortions can be evaluated by the point source image position on the detector as represented in Figure 5. Therefore, this is a useful measurement that provides abundant information.

On the other hand, the long-slit fore optics setup can also provide spectral resolution, smile distortion, and keystone distortion across the FOV except for spatial resolution. However, this measurement can be performed within one shot and there is no need to adjust the vertical stage. The schematic result of the long-slit fore optics setup is presented in Figure 6.

6. Optical Performance

6.1. Wavelength Registration Position Verification

The measured wavelength-registered position on the image plane is compared to the simulation results. This is achieved by extracting the centroid positions of some wavelengths in the HgAr spectrum from Zemax and comparing them with the positions of these wavelengths on the detector. The departure between the two results is presented in

Table 3. The wavelength registration comparison between theoretical (Sim.) and experimental (Exp.) positions.

Wavelength (nm)	Sim. Position (mm)	Exp. Position (mm)	Departure (µm)
0th	0	0	0
435.832	0.9423	0.9445 ± 2.6	2.2
546.073	1.1806	1.1825 ± 0	1.9
576.960	1.2474	1.2485 ± 0	1.1
696.543	1.5060	1.5070 ± 0	1.0
763.511	1.6508	1.6500 ± 0	0.8
810.369	1.7521	1.7545 ± 0	2.4
826.452	1.7869	1.7875 ± 0	0.6
840.821	1.8180	1.8220 ± 2.6	2.5
912.298	1.9726	1.9745 ± 0	1.9

. It is performed by setting a reference at the 0th diffraction order. It is important to mention that these results are measured at the center of the slit.

However, there is some mismatch to position according to the discreteposition of pixels. The position mismatch acceptable range depends on the pixel width. This is described by Figure 7. Therefore, the acceptable range is equal to half a pixel width, which is ±2.75 µm in our case. The evaluation proves that the Offner spectrometer dispersion performance agrees with the simulation as shown in Table 3 and that the position mismatch is within the acceptable range.

6.2. Spatial and Spectral Resolutions

The point-object LPU was used for this measurement. The measurement was performed at the center of the slit to establish a baseline for the system’s optimal performance, as optical aberrations and distortions are typically minimized at this location. The spatial and spectral resolutions can be evaluated by measuring the FWHM of the spot image [9]. An example of a FWHM measurement can be seen in Figure 8. The FWHM of the spot image measurement in the spatial direction determines the spatial resolution of the optical system, while the measurement in the spectral direction determines the spectral resolution of the optical system. In our case, 5.5 µm in spatial direction implies a 15 m resolution on the Earth’s surface. On the other hand, 5.5 µm in spectral direction implies 2.5 nm in the spectral domain. The FWHM of a point source image for 400 nm to 1000 nm spectral domain, in both spatial and spectral directions, are presented in

Table 4. The FWHM of a spot image in the spatial direction and the spectral direction for the simulation (sim.) and the experiment (exp.).

Wavelength (nm)	FWHM
	Spatial Direction (µm)		Spectral Direction (nm)
	sim.	exp.	sim.	exp.
405	21.3	21.5 ± 0.1	7.4	7.7 ± 0.1
505	21.0	21.0 ± 0.2	7.3	7.3 ± 0.1
595	21.6	21.6 ± 0.2	7.5	7.6 ± 0.1
700	22.5	22.6 ± 0.1	7.7	7.9 ± 0.1
810	23.1	23.2 ± 0.2	8.4	8.4 ± 0.1
880	23.6	23.8 ± 0.1	8.5	8.6 ± 0.1
970	23.9	24.1 ± 0.2	8.6	8.7 ± 0.1

. This shows that the simulation and the measurement results are in good agreement.

It is noticeable that the FWHM in the spatial direction varies with respect to wavelengths according to diffraction law as shown in Figure 8 left-panel. The FWHM varies from 21.5 µm to 24.10 µm in the spatial domain. However, it is visible that FWHM at 400 nm is higher than at 500 nm. This confirms the simulation on Zemax that the image quality at 400 nm is lower than at 500 nm [1].

On the other hand, the FWHM in the spectral direction varies between 7.3 nm and 8.6 nm in the spectral domain. It varies according to the diffraction law. Therefore, longer wavelengths usually achieve lower spectral resolution. This means the instrument can distinguish fine spectral features more effectively at shorter rather than longer wavelengths. Consequently, the spectral resolution is not uniform across the entire wavelength range but exhibits slight variations.

The FWHM measurement in both spatial and spectral directions confirms that the experimental results are very close to the simulation predictions. This close agreement indicates the high accuracy and reliability of the design, validating the effectiveness of the optical alignment and the overall system configuration.

6.3. Distortions

The distortion of the spectrometer can be measured by using either point-object fore-optics or long-slit fore-optics. First, we measured the distortion by using point-source fore optics. Moving the vertical stage and collecting the data, we obtained the data set across 27 points along the slit as presented in Figure 9. This is the data set from the HgAr source. The 0th diffraction order is perfectly designed to be at the center of the detector. The +1st diffraction order lies on the right side of the 0th order.

To measure smile distortion, we evaluated the straightness of the dispersed wavelength across FOVs. By comparing the location of the same wavelength across FOVs, we determined that the center of the 435.8 nm PSFs across FOVs is located at pixel column 1196. This is also observed for other wavelengths, such as the center of the PSFs across FOVs for wavelength 546.0 nm on pixel column 1239. Therefore, we can confidently say that the optical system is free from smile distortion. This can also be easily detected by the naked eye, as shown in Figure 10, since the smile distortion is usually invariant for various wavelengths. Measuring only a few wavelengths can confirm the existence of smile distortion. In our case, it is visible that the same dispersed wavelength across FOVs perfectly falls onto the same pixel row. Therefore, we can conclude that the TSC-1 Offner spectrometer is free from smile distortion.

Keystone distortion can be measured by evaluating the straightness of the spectrum in each FOV. It can be accomplished by comparing the position of various wavelengths in a FOV as presented in Figure 11. It is important to mention that the keystone is more observable at the edge of the FOV. Therefore, we aim to measure the keystone near the edges of the FOV. We first measured at the one edge. The center of the PSFs for various wavelengths is perfectly located at pixel row 55, which corresponds to position +5.340 mm on the slit. On the side close to the other edge, the center of the PSFs for various wavelengths is perfectly located at pixel row number 1861, which corresponds to position −4.588 mm on the slit. Moreover, the center of the PSFs for various wavelengths is perfectly located at pixel row 1024 for the center of the FOV. Keystone distortion can also be measured by using a long-slit LPU equipped with a tungsten calibration source that provides a continuous spectrum. The measurement result is presented in Figure 12. The naked eye can observe that the edges of the spectrum appear in a straight line, with sharp straight edges at the top and bottom of the image. The spectrum is cut at the pixel rows 25 and 2024. Therefore, we can confidently say that the Offner prototype is free from keystone distortion.

In conclusion, the performance testing conducted to evaluate the Offner spectrometer’s performance has yielded promising results. Smile distortion, which causes a curvature in the spectral lines, was not observed in the experimental data. Additionally, keystone distortion, characterized by the spectral lines’ tilt across the image plane, was also not evident in the measurements. Notably, the absence of smile and keystone distortions in the measured data reaffirms the accuracy and validity of the theoretical simulations, where the Offer is well known for its distortion-free results. This demonstrates that the optical design and alignment of the spectrometer have been well executed.

6.4. Signal-to-Noise Ratio

The signal-to-noise ratio (SNR) was evaluated by using the signal from an HgAr lamp spectrum measurement after dark subtraction. The odd number of dark frames, 11 in our case, were acquired with no light on at the same exposure time as the raw data. Then, they were stacked to average the dark signal for each pixel before subtraction. The measurement result for SNR evaluation is presented in Figure 13. The signal equals to 200 Analog-to-Digital Units (ADU) at 550 nm. On the other hand, the noise is evaluated by determining the variance in the no-data zone of the image, which includes the rows numbered 1–24 and 2025–2048, represented in Figure 12 and equal to 2 ADU. Therefore, the SNR is equal to 100 at 550 nm.

6.5. Specimen Spectrum Measurement

A leaf is used to collect a reflected specimen spectrum to access the use of this instrument. A tungsten source is used to illuminate the leaf, which is located on the flat screen in the long-slit LPU, as shown in Figure 14. The raw data cross-section at the center of the slit is presented in Figure 15 right-panel in the blue curve. Then, the data were calibrated by subtracting the dark frame, performing the wavelength calibration, and applying the sensor spectral response correction factors. The dark subtraction was achieved the same way as for the SNR evaluation. After that, the known emission peaks calibration source, including mercury–argon (HgAr) and neon (Ne) calibration lamps, are used to accurately convert pixel numbers to wavelengths. The HgAr lamp emission lines range from approximately 400 to 900 nm, except from 600 to 700 nm. An additional neon lamp is used to fill this gap. This process establishes the relationship between wavelength and pixel number. Once this relationship is determined, it is applied to the plot scale to accurately identify the corresponding wavelength for each pixel. Finally, the sensor spectral response is measured using a tungsten continuous spectrum calibration source and then compared with the tungsten spectrum reference data for the correction factors. The calibrated leaf data are presented in Figure 15 left-panel and the cross-section at the center of the slit is presented in Figure 15 right-panel in the orange curve. It is important to mention that these data are in the ADU scale. A reflectance calibration object is needed to achieve a reflectance value. However, the remarkable sharp absorption feature around 700 nm is known as the “red edge”. This feature is found in the reflectance spectrum of healthy green vegetation. The phenomenon is primarily caused by the strong absorption of chlorophyll in the red region of the spectrum and the relatively weak absorption in the near-infrared region [19].

7. Conclusions

The development of the TSC-1 Offner spectrometer prototype has been a significant milestone in Thailand’s creation of space optical instruments. The prototype development and experimental results are addressed and discussed, including some ongoingissues and solutions. The prototype performance and comparisons with simulation data affirm the spectrometer’s distortion-free imaging capabilities. The PSF FWHM in the spatial direction varies between 21.0 to 24.1 µm, which implies 57 to 66 m on the ground when equipped with a 231 mm focal length telescope. The successful construction and calibration of the prototype demonstrate the feasibility of the Offner spectrometer for practical use with 7.3 to 8.6 nm spectral resolution throughout the specified spectral domain. The SNR 100 is achieved at 550 nm for the HgAr spectrum. Overall, this work solidifies the foundation for further advancements and potential real-world deployment of the Offner spectrometer for Earth observation missions. The specified spectral range and the spectral resolution provide detailed spectral information for a wide range of Earth observation applications, including vegetation analysis [20], land cover classification [21], water quality assessment [22], atmospheric studies [23], and solar-induced fluorescence (SIF) remote sensing [24]. It enables researchers and scientists to derive meaningful insights about the Earth’s surface and dynamic processes.

In order to complete the TSC-1 hyperspectral imager system, we need a front telescope for which the design is already complete [1]. We plan to build the front telescope prototype right after we finish the characterization of the Offner prototype. The telescope’s mirrors will be made of aluminum and follow a single-material strategy. However, using an off-axis primary mirror poses a challenge in terms of alignment. Although the spectro-base used for the Offner prototype ensures that the mechanical reference is capable of optical alignment, the front telescope structure is much larger than the spectrometer unit and cannot be built in a single part. Therefore, we still need to investigate the proper instructions to build and align the front telescope for the off-axis telescope configuration. Then, the front telescope will be characterized individually. After that, the two parts will be combined and tested as a fully functional hyperspectral imaging system. Moreover, for potential future work, the performance metrics shall be assessed with detailed statistical methods, which leads to a more comprehensive interpretation and advantage for the scientific implications of Earth observation. In addition, more case studies and simulations would allow the use of the TSC-1 hyperspectral imager to improve the practical development of the space research roadmap in Thailand. Furthermore, we plan to conduct the prototype’s thermal characterization in the near future. We intend to report the results in a dedicated communication, once the thermal characterization experiments are completed.