Point Cloud Registration with data measured from a photon-counting LIDAR sensor from a large distance (500 m - 1.5 km) is an expanding field. Data measuredfrom far is sparse and have low detail, which can make the registration processdifficult, and registering this type of data is fairly unexplored. In recent years,machine learning for point cloud registration has been explored with promisingresults. This work compares the performance of the point cloud registration algorithm Iterative Closest Point with state-of-the-art algorithms, with data froma photon-counting LIDAR sensor. The data was provided by the Swedish Defense Research Agency (FOI). The chosen state-of-the-art algorithms were thenon-learning-based Fast Global Registration and learning-based D3Feat and SpinNet. The results indicated that all state-of-the-art algorithms achieve a substantial increase in performance compared to the Iterative Closest Point method. Allthe state-of-the-art algorithms utilize their calculated features to obtain bettercorrespondence points and therefore, can achieve higher performance in pointcloud registration. D3Feat performed point cloud registration with the highestaccuracy of all the state-of-the-art algorithms and ICP.

No description available.

Computed Tomography (CT) is a non-invasive x-ray imaging method capable of reconstructing highly detailed cross-sectional interior maps of an object. CT is used in a range of medical applications such as detection of skeletal fractures, organ trauma and artery calcification. Reconstructing CT images requires the use of a forward operator, which is essentially a simulation of the scanning process. Photon-Counting CT is a rapidly developing alternative to conventional CT that promises higher spatial resolution, more accurate material separation and more robust reconstructions. A major difficulty in Photon-Counting CT is to model cross-talk between detectors. One way is to incorporate a wide point-spread function into the forward operator. Although this method works, it drastically slows down the reconstruction process.  In this thesis, we accelerate image reconstruction tasks for photon-counting CT by approximating the cross-talk component of the forward operator with a deep neural network, resulting in a learned forward operator. The learned operator reduces reconstruction error by an order of magnitude at the cost of a 20% increase in computation time, compared to ignoring cross-talk altogether. Furthermore, it generalises well to both unseen data and unseen detector settings. Our results indicate that a learned forward operator is a suitable way of approximating the forward operator in photon-counting CT. / Datortomografi (CT) är en icke-invasiv röntgenmetod som kan skapa högupplösta tvärsnittsbilder av objekt. CT används i en stor mängd tillämpningar, exempelvis vid detektion av frakturer, mjukvävnadstrauma och åderförkalkning. När man rekonstuerar tvärsnitt i CT krävs en simuleringsmodell som kallas framåtoperatorn. Fotonräknande CT är ett alternativ till konventionell CT som utlovar högre upplösning, mer precis uppdelning av material och högre robusthet i rekonstruktionerna. I fotonräknande CT är det viktigt att ta hänsyn till överhörning mellan detektorerna. Ett sätt är att inkorporera en punktspridningsfunktion i framåtoperatorn, vilket dessvärre saktar ned rekonstruktionsprocessen drastiskt.  I detta examensarbete approximerar vi överhörningseffekten mellan detektorer med ett djupt neuralt nätverk, med syfte att accelerera rekonstruktionsprocessen för fotonräknande spektral CT. Den inlärda framåtoperatorn reducerar rekonstruktionsfelet med en faktor tio på bekostnad av en 20-procentig ökning i beräkningstid, jämfört med en framåtoperator som inte modellerar överhörning. Vi visar att den inlärda framåtoperatorn generaliserar väl till data som den inte är tränad på, men även detektorinställningar den inte är van vid. Våra resultat tyder på att den inlärda framåtoperatorn är en lämplig approximationsmetod för framåtoperatorn i fotonräknande CT.

Photon counting computed tomography scanners constitute a major improvement of the field of computed tomography, opening various prospective and enabling the decomposition of computed tomography images into different materials. The material decomposition algorithm, mapping photon counts to material pathlengths, relies on a forward model with Poisson statistics. This model though suffers from noise and residual bias due to its sensitivity to calibration errors and specificities in single-pixel responses that are not captured by the material decomposition model.           This study proposes a pixel-specific and projection-based correction of the residual bias in the material decomposition estimates using artificial neural networks trained for each pixel of the detector. The neural network models were trained under supervised learning using material decomposition calibration data, scans of PE and PVC slabs of various thicknesses acquired for the calibration of the model. This method aims at the mapping of the singularities of the pixels’ responses and correct them in the projection domain. The trained models were evaluated on a set of evaluation slabs and on scans of a water phantom, in order to assess performances of homogeneity and bias correction.           The implemented solution exhibited promising results for the correction of residual bias in single pixels without impairment of the noise levels. An array of trained neural networks demonstrates its ability to correct calibration and evaluation slab data while conserving pixel-to-pixel difference. The application of the correction to the water phantom however offered nuanced results which call for further investigation of the identified issues and induced improvements of the model.

The goal in medical x-ray imaging is to obtain the image quality requiredfor a given detection task, while ensuring that the patient dose is kept as lowas reasonably achievable. The two most common strategies for dose reductionare: optimizing incident x-ray beams and utilizing energy informationof transmitted beams with new detector techniques (spectral imaging). Inthis thesis, dose optimization schemes were investigated in two x-ray imagingsystems: digital mammography and computed tomography (CT). In digital mammography, the usefulness of anti-scatter grids was investigatedas a function of breast thickness with varying geometries and experimentalconditions. The general conclusion is that keeping the grid is optimalfor breasts thicker than 5 cm, whereas the dose can be reduced without a gridfor thinner breasts. A photon-counting silicon-strip detector developed for spectral mammographywas characterized using synchrotron radiation. Energy resolution, ΔE/Ein, was measured to vary between 0.11-0.23 in the energy range 15-40 keV, which is better than the energy resolution of 0.12-0.35 measured inthe state-of-the-art photon-counting mammography system. Pulse pileup hasshown little effect on energy resolution. In CT, the performance of a segmented silicon-strip detector developedfor spectral CT was evaluated and a theoretical comparison was made withthe state-of-the-art CT detector for some clinically relevant imaging tasks.The results indicate that the proposed photon-counting silicon CT detector issuperior to the state-of-the-art CT detector, especially for high-contrast andhigh-resolution imaging tasks. The beam quality was optimized for the proposed photon-counting spectralCT detector in two head imaging cases: non-enhanced imaging and Kedgeimaging. For non-enhanced imaging, a 120-kVp spectrum filtered by 2half value layer (HVL) copper (Z = 29) provides the best performance. Wheniodine is used in K-edge imaging, the optimal filter is 2 HVL iodine (Z = 53)and the optimal kVps are 60-75 kVp. In the case of gadolinium imaging, theradiation dose can be minimized at 120 kVp filtered by 2 HVL thulium (Z =69). / <p>QC 20160401</p>

This thesis explores the use of single-photon avalanche diodes (SPADs) for visible light communications (VLC). The high sensitivity of SPADs can potentially enhance the performance of VLC receivers. However, a SPAD-based system has challenges that need to be addressed before it can be considered as a viable option for VLC. The first challenge is the susceptibility of SPAD-based receivers to variations in ambient light. The high sensitivity of SPADs is advantageous for signal detection, but also makes SPADs vulnerable to variations in ambient light. In this thesis, the performance of a SPAD-based receiver is investigated under changing lighting conditions. Analytical expressions to quantify performance are derived, and an experiment is conducted to gain further understanding of system performance. It is shown that a SPAD-based receiver is highly sensitive to illumination changes when on-off keying (OOK) is employed, and that pulse-position modulation (PPM) is a preferred modulation scheme as it is more robust. The second challenge is broadcasting to SPAD-based receivers with different capabilities. A traditional broadcasting scheme is time-sharing, whereby a transmitter sends data to receivers in an alternating manner. Broadcasting to SPAD-based receivers is challenging as receivers may have diverse capabilities. In this thesis, a new multiresolution modulation scheme is proposed, which can potentially improve system performance over the traditional timesharing approach. The performance of the proposed scheme is analyzed, and a proof-of-concept experiment is performed to demonstrate its viability.

Dans le domaine de la tomographie par rayons X préclinique, la tomographie spectrale est une voie de plus en plus en plus explorée. Les objectifs de la tomographie spectrale sont la caractérisation et la quantification des tissus et agents de contraste que l'amélioration de contraste entre tissus mous. Cela passe par l'exploitation de l'information spectrale (ou énergétique) des photons X et non plus seulement par leur quantité détectée comme en tomographie standard par rayons X. L'intérêt de la tomographie spectrale se trouve renforcé par l'arrivée des caméras à pixels hybrides comme le XPAD qui ont la capacité de sélectionner les photons X en fonction de leur énergie. La caméra XPAD3, la troisième version du XPAD est construite pour fonctionner dans le micro-tomodensitomètre, PIXSCAN développé au CPPM.Dans ce contexte, cette thèse a deux buts~: participer au développement du PIXSCAN et effectuer une preuve de concept de la tomographie spectrale à l'aide du PIXSCAN. Le premier but est rempli grâce au développement de l'interface d'acquisition du PIXSCAN. Le second est accompli par l'implantation de la méthode par séparation de composantes dont le but est d'isoler les contributions photoélectrique, Compton et des agents de contraste. Ce travail débute par la caractérisation de la méthode et se termine par la preuve de concept sur données réelles acquises à l'aide du PIXSCAN. / In the field of preclinical X-ray tomography, spectral tomography is actively explorated. The aims of spectral tomography are the caracterisation of tissues and contrast agentstogether with the quantification of the latter and the enhancement of contrast between soft tissues. This is achived by the exploitation of spectral information (i.e. energy) and not only the detected quantities of photons X. The interest in spectral tomography is enforced by the arrival of hybrid pixel cameras like XPAD, because of their ability to select photons according to their energy. The XPAD3 camera, third version of XPAD, is built to be used in the micro-CT demonstrator PIXCAN fully developped at CPPM.In this context, this thesis has two goals : a contribution to the developement of the PIXSCAN and a realisation of a proof of concept of spectral tomography in PIXSCAN. The first goal is done by developing the data acquisition system of PIXSCAN. To accomplish the second one, we perform spectral tomography by implementing component separation in order to isolate photoelectric, compton and contrast agents contribution. This work begins by the caracterisation of this method and ends by a proof of concept on real data acquired by PIXSCAN.

La tomographie à rayons X est une technologie d’imagerie en trois dimensions. Elle se base sur la transmission de rayons X à travers l’objet d’étude. Elle est non destructive mais néanmoins irradiante. Cette technique de visualisation est utilisée principalement dans trois domaines : le diagnostic médical, le contrôle non destructif (détection de défauts dans des pièces industrielles de haute performance) et la sécurité (contrôles aéroportuaires des bagages). Les récentes avancées technologiques dans le domaine des détecteurs spectrométriques de rayons X ouvrent des perspectives d’amélioration de cette technique d’imagerie dans ses divers domaines d’application. Nous avons développé une nouvelle méthode reconstruction statistique appelée MLTR-ONE-STEP qui permet de reconstruire la variabilité énergétique du coefficient linéaire d’atténuation de l’objet étudié. Cette approche est dite « one-step » car elle reconstruit directement le volume final à partir des mesures brutes issues de détecteurs spectrométriques.Les phénomènes physiques au sein du détecteur provoquent une distorsion énergétique du spectre d’atténuation qui a été prise en compte lors de la reconstruction. La méthode utilisée s’inscrit dans le cadre bayésien et maximise la log-vraisemblance du modèle tout en prenant en compte de l’a priori spatial sur le volume reconstruit. L’objectif de la méthode est l’amélioration de la qualité de l’image finale (réduction des artefacts et niveau de bruit) et la quantification des matériaux présents. Nous avons étudié dans le cadre de données simulées l’influence des paramètres de régularisation sur la reconstruction. En pratique, le détecteur de rayon X étudié classe les photons incidents en 64 canaux. Ils sont ensuite regroupés en un nombre de canaux plus faible (2 à 25) et l’influence de ce regroupement a été étudiée. La reconstruction MLTR-ONE-STEP a ensuite été testée sur des données expérimentales regroupées en 12 canaux. / X-ray spectral tomography is a 3D visualization technique. It is based on the transmission of X-rays through object matter. It is a non-destructive technology but which irradiates the studied object/patient. X-ray tomography is mainly used in three areas: medical diagnosis, non-destructive testing (detection of defects in industry devices) and airport security (luggage screening). New technological breakthroughs in X-ray photon-counting detectors provide new perspective for improving this technique in each application field. We have developed a new reconstruction method named MLTR-ONE-STEP which enables the obtention of energetic variability of the scanned object linear attenuation coefficient. This approach belongs to the “One-Step” class because it directly reconstructs the final images from raw photon-counting detector data.Physical effects inside the detector are causing spectral distortion of the energetic spectrum. This distortion is taken into account in our reconstruction through a Detector Response Matrix. The developed reconstruction method maximizes the poissonian likelihood of the measurements with a spatial regularization Tukey term. The objectives of spectral tomography are the improvement of the image quality compared to standard tomography and the quantification of materials inside the object. We have studied the influence of regularization parameters on the final result. In practice, photon-counting detector measurements are in practice sorted in 64 energy bins. Bins are then merged in a smaller number (from 2 to 25). The influence of this binning was studied on simulated data. The MLTR-ONE-STEP was then tested on real experimental data in order to prove the feasibility of such a “One-Step” reconstruction method.

Les interféromètres stellaires sont des dispositifs à très haute résolution angulaire, permettant une étude approfondie de l’Univers. Cette thèse décrit la mise en place en laboratoire d’un interféromètre à somme de fréquences dédié à la détection de rayonnement infrarouge en bande L :ALOHA@3.39. Afin de pouvoir limiter le bruit généré par les rayonnements thermiques ambiants et de pouvoir bénéficier des technologies matures en terme de détection et de transport cohérents de la lumière collectée, ce système intègre sur chacune des voies interférométrique un processus non linéaire de somme de fréquence. Les rayonnements autour de 3.39 μm sont transposés autour de 810 nm via des cristaux de PPLN alimentés par une pompe unique à 1064 nm. Une première partie présente le contexte expérimental et théorique de la thèse. La deuxième partie concerne la conception, la réalisation et la caractérisation du banc ALOHA@3.39. La troisième partie présente les résultats expérimentaux obtenus en laboratoire. La mesure répétée du contraste en régime de fort flux permet de calibrer le contraste instrumental du montage. Les premières mesures en régime de comptage de photons montrent que nous sommes actuellement capables de détecter des franges d’interférence avec de hauts contrastes en ne disposant que de 100 Fw à l’entrée de chacune des voies. Enfin, nos investigations nous ont amené à la détection de franges d’interférences via le prototype ALOHA@3.39 à partir d’une source thermique, assimilable à un corps noir. Cette thèse conclue sur une estimation de la magnitude limite accessible, et sur l’utilisation de nouvelles technologies de PPLN. / Stellar interferometers are high angular resolution devices, allowing for detailed research of the Universe. This thesis describes the in-lab implementation of a sum frequency generation interferometer dedicated to the infrared detection in the L band : ALOHA@3.39. In the aim of limiting thermal noise due to the room radiation, and benefit mature detectors and fibered components, this device includes a sum frequency generation non-linear process on each of its arms. Stellar radiations around 3.39 μm are transposed to around 810 nm thanks to PPLN cristals powered by a single pump signal at 1064 nm. The first part presents our global experimental context and theoretical elements about this thesis. The second part deals with the conception, the implementation and characterisation of the test bench ALOHA@3.39. The third part tables in-lab experimental results. Repeated measurements of the fringe contrast on high flux regime allow to calibrate the instrumental contrast of the set-up. First measurements on photon counting regime show we are currently able to detect interference fringes with high contrast with only 100 fW at the input of each arm. Finally, our research led us to realise interference fringes detection via the ALOHA@3.39 prototype, observing a thermal source, considered as a black body. This thesis concludes on an estimation of the limiting reachable magnitude, and on the future use of new PPLN technologies.

Le processus de somme de fréquences optiques est utilisé dans certaines applications pour convertir des signaux de longueurs d’onde infrarouges vers le domaine de longueurs d’onde visibles. Cela permet de bénéficier de technologies plus performantes notamment en terme de détection et de propagation. Les travaux menés dans cette thèse s’intéressent à l’étude de phénomènes optiques parasites générés par ce processus non linéaire dans des cristaux de niobate de lithium périodiquement polarisés (PPLN) pour des applications de conversion de fréquences en régime de comptage de photons. La première partie de ce manuscrit montre l’intérêt du processus de somme de fréquences optiques dans le contexte technologique de la détection infrarouge en régime de comptage de photons, via le concept de détection hybride. Les éléments théoriques et l’état de l’art associés à la détection hybride seront également présentés dans cette première partie. La deuxième partie traite d’une étude comparative entre différents cristaux de PPLN pour la détection hybride à 1550 nm. Pour cela, une analyse expérimentale détaillée des processus parasites est effectuée. Les résultats de cette étude sont utilisés pour des applications en astronomie et en microscopie. Enfin, la troisième partie traite d’une application de la somme de fréquences en interférométrie stellaire. Les résultats de caractérisation des cristaux de PPLN sont mis à contribution dans l’optimisation d’un instrument pour l’astronomie appelé interféromètre à somme de fréquences. Les résultats d’observation sur le site astronomique du Mont Wilson ainsi que les perspectives de cette instrument y sont présentés. / The sum frequency generation process is used in differents applications to convert signals from infrared wavelengths to the field of visible wavelength. This allows to benefit of more efficient technologies in terms of detection and propagation. This thesis describes the study of noise phenomena generated by this process in crystals of periodically poled lithium niobate (PPLN) for frequency up-conversion applications in photon counting regime. The first part of the manuscript shows the advantage of sum frequency generation process in the technological environment of the infrared detection on single photon counting regime, using the concept of hybrid detection. The theoretical elements and the state of the art associated with hybrid detection will also be presented in this first part. The second part deals with a comparative study between different PPLN for hybrid detection at 1550 nm. A detailed experimental analysis of the noise process is performed. The results of this study are used for applications in astronomy and microscopy. The third part deals with an application of the sum frequency in stellar interferometry. The results of PPLN’s characterization are involved in optimizing an instrument for astronomy called sum frequency interferometer. The results of observation on the astronomical site of Mount Wilson and the prospects of this instrument are presented.