We present the modeling and the performance of a polarization
active imager at l5806 nm. The device operates in a monostatic
configuration, using a semiconductor laser to illuminate the target and a
telescope to create the image on a CCD matrix. Dual images (intensity
and polarization degree) of different scenes are obtained by a new
method (only two images acquired) and analyzed, showing the experimental
validation of this concept. An application of this active imager to
the detection of a target buried in the background (same reflectivity but
different polarization degree) is proposed. Field experiment results are
reported.

Modeling and performances of a polarization active imager at λ=806 nm

We present the modeling and the performances of a Polarization Active Imager at (lambda) equals 806 nm. The device is operating in a monostatic configuration, using a semiconductor laser to illuminate the target and a telescope to create the image on a CCD matrix. Dual images (intensity and polarization degree) of different scenes are obtained by a new method (only 2 images acquisition) and analyzed, showing the experimental validation of this concept. The application of this active imager to the detection of target buried in the background (same reflectivity but different polarization degree) is proposed.

Modeling and performance of a polarization active imager at λ=806 nm

In the automatic detection for surface defects of optical components, the digs and dust particles exhibit similar features: point-like shape and variable intensity reflectivity. On this condition, these two types with entirely different damages are easily confused so that misjudgments will be induced. To solve this problem, a polarization-characteristics-based classification method of digs and dust particles (PCCDD) is proposed based on the polarimetric imaging technique and dark-field imaging technique. First, a dark-field imaging system equipped with a polarization state generator (PSG) and a polarization state analyzer (PSA) is employed to measure and establish normalized Mueller matrices' datasets of digs and dust particles. And by a nonlinear global search combined with a separability evaluation method, the optimal number of acquisitions and corresponding polarization measurement states of the PSG and the PSA are obtained, as well as the parameters of classification function. Then, multiple polarization images are acquired under the optimal states to extract a multidimensional feature description that relates only to the polarization characteristics of the defect; this subsequently acts as the input vector of the classifier to finally achieve the classification. This method takes full advantage of both the difference in polarization properties between digs and dust particles and the characteristic that the polarization properties of digs are relatively invariant while those of dust particles have a large variability. The classification process involves only simple matrix operations. Compared to the traditional discrimination method based on intensity images, the features obtained by this method have a higher separability. Experiments show that the classification accuracy reaches over 90%. This method can be further applied to the recognition and discrimination of other defects in the field of surface defects' detection.

Classification between digs and dust particles on optical surfaces with acquisition and analysis of polarization characteristics.

Ce travail concerne les developpements necessaires a la mise en oeuvre efficace de systemes imageurs polarimetriques et comporte plusieurs volets allant de la theorie amont a l'utilisation d'algorithmes de traitement d'images specifiques. Les principaux objectifs et les developpements realises se rapportent a: - la representation algebrique des formalismes de polarisation en utilisant les quaternions : toute matrice de Mueller non depolarisante peut etre mise sous forme d'un quaternion complexe et l'action d'un systeme optique sur une onde monochromatique s'ecrit comme le produit geometrique du quaternion correspondant a la matrice de Mueller par le multivecteur representant le vecteur de Stokes. Cette utilisation de l'algebre geometrique permet de definir formellement la notion de distance, de similitude et d'appartenance a une classe. - l'optimisation des polarimetres imageurs et le calcul de l'erreur sur les mesures : la demarche, basee sur l'utilisation du produit de Kronecker, prend en compte de facon conjointe les erreurs sytematiques et le bruit d'image et leur minimisation est obtenue des l'etape de conception. Plusieurs fonctions de merites sont introduites pour permettre la definition des caracteristiques du polarimetre optimal. - l'estimation et le filtrage du bruit des images polarimetriques : deux methodes sont envisagees i) une nouvelle variante de la methode du « scatter plot » combinee avec une vectorisation de l'image par un parcours de type Piano-Hilbert, ii) la methode du masquage de donnees qui repose sur l'utilisation de la difference entre deux approximations de l'operateur Laplacien. Les performances et les biais des deux estimateurs choisis sont etudies statistiquement par la methode du « Bootstrap ». - la classification et la previsualisation couleur des images codees en polarisation : on propose une representation coloree des images codees en polarisation, comme une aide a leur interpretation en fonction de leur contenu physique, qui utilise la decomposition polaire pour le cas de l'imagerie de Mueller et repose sur deux mappages entre la sphere de Poincare et un espace de couleur parametrique dans le cas de l'imagerie de Stokes. Le processus de segmentation est base sur la famille des algorithmes des K-moyennes. Cette demarche est illustree sur des images de Stokes et de Mueller de tissus biologiques.

https://tel.archives-ouvertes.fr/tel-00443685/document

Analyse et traitement des images codées en polarisation

In a recently developed compact LiDAR polarimeter, the transmit beam cycles through multiple polarization states within each laser pulse and the receiver splits the received signal into multiple polarization state analyzers (PSAs), with the PSA outputs temporally multiplexed into a single detector. This enables measurement of up to 12 of the 16 Mueller matrix elements of a target using limited hardware. However, due to numerical issues, one entire column of the matrix is not accessible, including the M22 element (counting from zero) on the matrix diagonal. Experimental data show that for most surfaces of interest in a defense/security setting, the off-diagonal elements tend to be negligibly small and the diagonal elements vary between targets, so access to the diagonal elements is of high interest. In this paper, we show that if an elliptical PSA is used at the receiver and if most of the off-diagonal Mueller matrix elements are assumed to be zero a priori, then the otherwise inaccessible M22 element can be estimated. We explore the linear algebra of the problem to determine the full list of which subsets of Mueller matrix elements can be estimated with this hardware configuration. We also theoretically and empirically investigate the effects of the rotations of the electro-optic transmitter plate and receiver elliptical PSA on the overall system performance.

Compact lidar polarimetry via time-varying transmit polarization and an elliptical polarization analyzer

A new bistatic carbon-dioxide laser reflection polarimeter has been developed. Measurements were done for a variety of soils and particulate media samples, Mueller matrix extracted and analysis are demonstrated. Dominant scattering mechanism is found to be a Fresnel type local interaction. A fully polarimetric radiative transfer model is developed, with phase matrix appropriate for locally specular reflection. The model is compared with previous absolute measurements of bistatic polarized scattering cross section. Optical constants, sample filling factor and incoherent scattering term are extracted. Excellent agreement is found between measured and modeled quantities.

Marc-Philippe Stoll

Papers

Laser polarimetry at 10.6 μm and modeling of scattering by soils and particulate media

Stokes images analysis using Poincaré sphere to color space mapping