Mission
1) Work environment
The internship will take place at the CNES site in Toulouse within the Image Quality team, which is responsible for the image quality of planetary and Earth observation cameras and the associated processing pipelines. The trainee will work closely with the team’s postdoctoral researcher, who established the radiometric and angular correction for the CAM8 instrument. The geological aspects of the study will be conducted in collaboration with the planetary geology team at CRPG (CNRS, Nancy).
2) Context
The Mohammed Bin Rashid Space Centre’s RASHID 3 rover is scheduled to explore the lunar surface no earlier than 2029. Building on the flight experience of the CASPEX camera suite flown on RASHID 1 [1], CNES is developing the imaging payloads and their processing pipelines for this mission. Among these payloads, CAM8 is a snapshot spectral camera based on a CMV4000 image sensor equipped with a monolithic mosaic of Fabry-Pérot interference filters. Each 4 × 4 pixel pattern combines eight narrow spectral bands with eight panchromatic pixels. With bands distributed between roughly 550 and 940nm, the system records a VISNIR spectrum across the entire field of view in a single exposure, without relying on filter wheels or moving parts. This spectral domain encompasses critical absorption features for lunar mineralogy. Iron oxides exhibit a steep slope and an absorption shoulder between 500 and 650nm, alongside a distinct band near 860nm. Concurrently, pyroxenes produce a broad absorption band centered between 900 and 1000nm [6]. These signatures are key to discriminating the main rock types and regolith components on the Moon.
Typical rover cameras are equipped with wide angle optical lens assemblies, which produce high incidence angles on the sensor. A well-known physical property of Fabry-Pérot filters is their angular sensitivity. Systematic measurements performed at CNES and CEA-Leti have quantified this effect on the CAM8 sensor [2]. On band B6, for instance, the transmission peak shifts 34nm toward shorter wavelengths at an incidence of 30 degrees, while peak amplitude varies by 16.5% between opposite angles of equal value. In addition, out-of-band transmission rises from 22% 1 at normal incidence to 43.5% at 30 degrees. Without appropriate correction, a given mineral yields disparate spectral signatures depending on its location between the center and the periphery of the image. The postdoctoral research work within the team has yielded an inversion model that corrects for these effects, outputting calibrated physical data in radiance and surface reflectance.
The objective is now to translate these calibrated physical products into mineralogical information. In planetary exploration, this is commonly achieved through spectral criteria, which combine bands using simple mathematical operations such as ratios, slopes, or continuum removed band depths. For example, the CRISM science team formulated sixty such parameters for Mars surface analysis [4], and an analogous framework was evaluated using a terrestrial emulator of the ExoMars PanCam instrument [5]. These studies underscore two essential con straints: while a criterion provides diagnostic evidence of a phase, it does not constitute unique identification, and its reliability depends directly on radiometric accuracy [3, 5]. A field campaign conducted in November 2025 in the Atacama Desert provided a rich multi spectral dataset over terrestrial analogues of lunar and Martian surfaces, serving as the primary benchmark for this internship. If needed, simulations or another dataset can complement it.
3)Objectives
The objective is to design, implement, and validate spectral criteria tailored to the eight bands of CAM8 to support mineralogical mapping of the observed landscapes. The work will proceed along six steps:
1. Review the literature. The trainee will study the spectral criteria used by planetary missions such as CRISM and PanCam [4, 5], and the absorption features of the target minerals in the CAM8 spectral range.
2. Understand the data. The trainee will get familiar with the CAM8 operating princi ples, take ownership of the Atacama dataset alongside the corrected radiance and reflectance products, and evaluate inter-band consistency across uniform reference areas.
3. Simulate CAM8spectra. Thetraineewill convolve mineral spectra from reference libraries (such as RELABandUSGS)aswell as available field measurements with the measured CAM8 responses to determine which phases can be reliably separated.
4. Define the criteria. Adapting the formulations used by CRISM and PanCam [4, 5], the trainee will design band ratios, spectral slopes, and band depths matching the CAM8 chan nels. The selected criteria and target phases, primarily pyroxenes and iron oxides, will be decided in coordination with the CRPG geologists.
5. Apply and validate. The trainee will generate parameter maps and color composite views from the Atacama data to cross-check results against geological ground truth. Robustness will be evaluated against noise, illumination geometry, and angular correction accuracy, specif ically comparing criteria computed before and after correction to measure the performance gain.
6. Deliver the tools. The resulting code will be implemented in Python, documented, and prepared for integration into the CAM8 ground processing pipeline. A final report will document the criteria, their operational boundaries, and recommendations for their use on the Moon.
[1] N. Théret, E. Cucchetti, E. Robert, C. Virmontois, et al., “Enhanced image processing for the CASPEX cameras onboard the Rashid 1 rover,” Space Science Reviews, vol. 220, no. 5, p. 60, 2024.
[2] A. Fsian, E. Robert, H. Dewitte, J. B. Thomas, N. Th´eret, and C. Virmontois, “Angular spectral characterization of a snapshot spectral imager based on filter arrays,” Color and Imaging Conference, 2026, under revision.
[3] R. J. Reid, P. H. Smith, M. Lemmon, et al., “Imager for Mars Pathfinder (IMP) image calibration,” Journal of Geophysical Research Planets, vol. 104, no. E4, pp. 8907 to 8925, 1999.
[4] Martinot, M., Besse, S., Flahaut, J., Quantin-Nataf, C., Lozac’h, L., & van Westrenen, W. (2018). Mineralogical diversity and geology of Humboldt crater derived using Moon Mineralogy Mapper data. Journal of Geophysical Research: Planets, 123, 612–629. https://doi.org/10.1002/2017JE005435
[5] E. J. Allender, C. R. Cousins, M. D. Gunn, and C. M. Caudill, “Multiscale and multispectral char acterization of mineralogy with the ExoMars 2022 rover remote sensing payload,” Earth and Space Science, vol. 7, e2019EA000692, 2020.
[6] J. B. Adams, “Visible and near infrared diffuse reflectance spectra of pyroxenes as applied to remote sensing of solid objects in the solar system,” Journal of Geophysical Research, vol. 79, no. 32, pp. 4829 to 4836, 1974.
Profil
Required curriculum: Master 2 or final year of engineering school in physics, remote sensing, image processing, or geosciences
Appreciated skills: Interest in spectroscopy and planetary science; basic knowledge of mineralogy is appreciated but not mandatory
Software: Python with NumPy, SciPy and Matplotlib