HERSCHEL/HECOR mission
successful launch 14/09/2009
>> more infos
PXRMS conference 14-18 feb.
BigSky - Montana - USA
>> more infos
SPIM Research
__________________________________________________________________________________
FOURIER TRANSFORM-BASED HYPERSPECTRAL IMAGING
Spectral imaging is the acquisition of the image of a scene in a large number of spectral bands. Since the currently limited size of the focal planes does not allow the simultaneous acquisition of the whole information, most instruments use a scanning device, this latter being either spectral or spatial. Unfortunately, this leads to using only a small amount of all the light available. Indeed, in the midwave and longwave infrared (3μm-5μm and 8μm-12μm), the detector noise is too significant to suffer such waste of light. A high étendue imaging static Fourier-transform spectrometer, incorporating a two-dimensional imaging system and a two-waves interferometer, is one solution to this issue, and is furthermore well suited to airborne imagery. We have designed such a system, focusing on its radiometric performances, its size, and the impact of optical aberrations. A demonstration instrument named CaHyd has been built up for the visible range (PhD of Yann Ferrec – DGA, see figure above). These studies have been conducted with our partner ONERA through the collaboration platform named PRECISION. Over 50 spectral bands in the visible were successfully acquired and processed.
This research is a follow-up to the previous work. Its objective is to model the physical and technological limits that inflence the quality of multi and hyper spectral images in the infrared domain. This study is conducted with an industrial partner and the ONERA.
__________________________________________________________________________________
POLARIMETRIC IMAGING
Polarimetric imaging consists in measuring the polarization state of the light that comes from any point of a scene. It can reveal contrasts that do not appear in classical images and makes it possible to remotely measure physical properties of the objects such as diffusion or ruggedness. Many types of polarimetric imaging systems exist. The simplest ones measure a single scalar property of state of polarization (degree of polarization, ellipticity…) while others measure the whole Stokes vector or even the Mueller matrix. This domain is rapidly growing at the international level, and the French community is particularly active. Indeed, F. Goudail is one of the main organizers of the annual national workshop on the topic.
During the past years, our team has studied simple polarimetric imaging systems that measure the orthogonal state contrast (OSC). They consist in illuminating the scene with a totally polarized beam and in forming two images: the first one (X) is formed with the light backscattered in the same state of polarization as the incident light. The second one (Y) is formed with the light polarized orthogonally to the incident state.These systems only require the acquisition of two images and are simple to operate; they can thus be added to existing mono-spectral or multi-spectral imaging systems to enhance their capacities. The first issue in such systems is to acquire the two images X and Y in a single shot, in order to measure rapidly evolving phenomena. For that purpose, we have designed and optimized a custom-based imaging architecture based on a polarization splitting Wollaston prism.
The second issue is estimation of the noise and optimisation of information extraction. Indeed, the OSC image is a ratio of images, and this operation enhances the noise that is present in intensity images X and Y. The resulting effect depends on the level and on the noise statistics. We studied the following cases:
• Additive Gaussian noise
• Non-uniform intensity of illumination.
• Simultaneous presence of signal-dependent and background Poisson noises.In all these cases, we have determined the estimation precision of the OSC (Cramer-Rao Lower Bound and variance of actual estimators). The results of these studies make it possible to assess more precisely the efficiency of OSC imagers and their added value compared to standard intensity imagers.
This research topic is recent in the team and opens many perspectives, both applicative and theoretical. In the future, we will investigate the use of OSC in conjunction with spectral imaging for characterizing diffusive materials and imaging through turbid media. We will also work at a rigorous definition of a contrast in OSC images. This approach will be generalized to the optimization of more complex polarization imaging architectures such as Stokes of Mueller imagers.__________________________________________________________________________________
WAVEFRONT ENGINEERING
The domain of wavefront engineering is rapidly growing. It consists in forming images with means other than standard refractive optics. The effect sought is performance enhancement of some sort, e.g., depth of field increase, size shrinking, or weight and/or cost reduction. These new systems are possible thanks to progress in microtechnologies, that make it possible to reliably fabricate tiny modulating structures, and to the fact that more and more computation power is available in imaging systems. Indeed, images from these new systems often require preprocessing or restoration before being presented to a visual observed or to an image analysis algorithm. This topic is currently very hot in optical engineering, and the main issues are:
• Phase mask technologies
• Optical design of imaging architectures
• Image processing for information extraction
We have built partnerships to address those different domains. The ultimate goal is to acquire an expertise on the global design of such systems.
a) Design of Pixellated Optics
Versatility is an important characteristic for phase modulation technologies since it makes it possible to adapt to a variety of aberrated wavefronts. Pixellated structures are good candidates if each pixel can be controlled independently. Such structures can be static or electrically monitored. On the other hand, spatial pixellation and phase quantification create image defects. Moreover, since pixels are small, their electromagnetic behaviour may not be well represented by standard Fourier optics. We are working with an industrial partner on developing that technology for ophthalmic and instrumental optics. Our purpose is to model the imaging properties of pixellated structures and to propose improvements to optimize image quality.
• We have studied the limits of scalar optics to model diffraction by the pixel structures. For that purpose, we have used a RCWA electromagnetic code developed by the NaPhEl group.
• We have studied the influence of the shape of the walls between the pixels, and proposed ways to reduce chromatic effects.
• We have studied spatial sampling strategies to optimize the quality of images provided by theses structures.
This work has been performed through two PhD students financed by the partner (G. Moulin, C. Pasanau).b) Aperture coding
In many applications, the space available for the imaging system is a crucial issue. Wavefront engineering is a good candidate for reducing their size. These studies are being developed in collaboration with ONERA with a PhD student financed by DGA and an industrial partner.
For example, we have shown the capabilities of an axicon for image-zooming using the well known great depth of focus of such components.
