315 Charact. and uses of cameras with Multi Spectral Filter Arrays (MSFA)

Job title:

315 Charact. and uses of cameras with Multi Spectral Filter Arrays (MSFA)

Company:

Centre National d’Etudes Spatiales

Job description

25-315 Charact. and uses of cameras with Multi Spectral Filter Arrays (MSFA)Postuler25-315 Charact. and uses of cameras with Multi Spectral Filter Arrays (MSFA)

• Post-doctorat, 24 mois
• Temps plein
• Moins de 2 ans d’expérience
• Doctorat, Bac+8
• Physical principles and image quality

PostulerMissionYour application must include a recommendation letter from your Ph.D. supervisor, a detailed CV including university education and work experience, a list of publications, a 2-page description of the work undertaken during the course of your PhD.For more Information, contact: Emilie ROBERT, (CNES technical referent)Submit the complete application online (Apply) before March 14th, 2025 Midnight Paris timeSpace exploration imaging serves contextual visualization for rover driving, understanding the geology of the visited soils, or light and atmosphere survey. Currently most of these images are spectral and done either through image sensors equipped with Color Filter Arrays (CFA) or through multiple images taken at different wavelengths selected by interferential filters on a rotating wheel. This last solution has the advantage of high spatial resolution, and has disadvantages. First, as a mechanism, it can be seen as likely to malfunction, in addition, it is very costly in terms of data transmission as an image per wavelength of interest as to be conveyed and a scene acquisition is very time consuming. Therefore, Multi-Spectral Filter Arrays (MSFAs), as a one-shot spatio-spectral sampling concept, appear as a dedicated solution. However, this cutting-edge technology and its uses through data processing need to mature to ensure a worthwhile deployment.TECHNICAL ENVIRONNEMENTThe concept of MSFA appeared in the early 2000s followed by simulation studies to help define the best designs. Then prototype trials began in the late 2010s and are still ongoing. The French Space Agency (CNES) is part of the early tryers of this technology. From the simulation period, several works brought to the conclusions that MSFA patterns should fulfill three requirements: (i) spatial uniformity to ensure robustness against image sensor defects; (ii) the designed pattern should be regular to ensure image reconstruction efficiency; (iii) the designed pattern should avoid the inconsistent optical crosstalk between adjacent pixels in imposing each wavelength-channel to have the same neighbor (neighbor-consistency). MSFA instruments despite major advantages therefore also come with manufacturing challenges linked to the deposition of interferential filter at pixel size always smaller, and to a fabrication process highly pattern- and wavelength- specific. In addition comes loss of incident signal and so decreasing of SNR in the image due to wavelength selection through filtering. CNES interest for this technology is linked to space exploration missions on the Moon (short-term) and/or Mars (mid-term).THIS PROJECT FOCUSThis postdoctoral project will focus on two main axis: (1) measurement of the spectral image sensor sensitivities, and (2) absolute calibration of the image signal to help determine the physical properties of planetary surfaces. Spectral sensitivities of image sensors equipped with MSFAs are highly dependent on the incident angle of the incoming light used for characterization at different wavelengths. Standard methods for measuring spectral sensitivities of Bayer image sensors make use of monochromatic flat fields for computing a mean signal over the whole matrix of pixels, or for each pixel. Studies conducted in collaboration between CNES and CNRS during a past postdoc on MSFAs have shown that such a method gives results for MSFA that are not representative of the spectral sensitivities of the imaging system during real-life acquisitions (with an optical system). It could be due to the impact of the optical system and therefore be worsened when wide-angle optics are used. A new method needs to be developped for characterizing spectral sensivities of such pixel matrices. This will also allow caracterisation of the cross-talk leading to signal mixing between different pixels and therefore spectral rejections. Note that the wavelength range of interest here is visible + near-infrared (VISNIR).The ultimate data need for geologists is spectral reflectance of the observed scene/landscape. A few works have been conducted to conclude which shape of spectral sensitivities is optimized for absolute calibration or for visualization. It shows it is really specific and can hardly satisfy multiple dedicated applications. The aim in this postdoctoral research project is to get the best from currently existing MSFA image sensors in optimizing the algorithms of the image processing pipeline. It therefore deals with spectral inversion using the spectral sensitivities measured during the project 1st part, and requires spectral rejection correction as well as spectral deconvolution. The methods might depends on the physical properties of the scene to characterize, therefore the project might need to begin with discussion with the science team of a planned exploration mission in order to better understand and take ownership of their needs.The results and perspectives of this project will help improve future exploration missions through:

• the definition of dedicated acquisition modes (which would allow in-situ calibration and data for instance for super resolution or HDR)

• New instruments architecture

• New specifications and methods for in-ground characterization of flight models

• New image processing pipelines allowing a more effective data usage

ProfilThe 1st axis of the postdoctoral project is practical, and requiers to be familiar (if not, be willing to learn) with image sensor opto-electrical characterization, and the developpement of setups. The second axis of the postdoctoral project requiers to be familiar with applied mathematics, and coding algorithms. It would be a plus to have some knowledge in radiometric models and optical properties of materials.

Expected salary

Location

Toulouse

Job date

Wed, 05 Feb 2025 06:06:37 GMT

To help us track our recruitment effort, please indicate in your email/cover letter where (vacanciesin.eu) you saw this job posting.