Search

Capturing and preprocessing 3D data is an increasingly important field, with advances in 3D displays, virtual reality, and digital fabrication driving the need for high-quality 3D data. Systems capable of acquiring 3D content can be separated into two classes: active and passive. Time-of-Flight (ToF) cameras are popular active depth sensors that suffer from significant noise when reconstructing geometry. In the first half of this thesis, we show a novel two-stage noise removal strategy for multi-view ToF systems in which the depth maps are first processed by a Convolutional Neural Network (CNN) to remove most of the depth map noise, and then further processed by a special 3D neural network to resolve inter-camera inconsistencies. Results on real data show a systematic improvement over noise removal tools from the literature. We then present a calibration method for accurately estimating the pose of multiple ToF cameras by using a 3D calibration object, ensuring proper alignment of the reprojected point clouds over multiple positions within the scene's 3D volume. In the second half of this thesis, Light Field (LF) systems are presented as passive depth sensors. A new geometric model and calibration method for plenoptic 1.0 devices is showcased, describing distortion effects at the microlens level for the first time. Improved quality in 3D reconstruction shows the validity of the model. A 3D Convolutional Neural Network is then shown for depth estimation from an LF image array, allowing for faster and lighter training than regular CNN from the literature. Finally, a novel spherical LF acquisition system is presented, capable of placing a camera at arbitrary positions around an object with promising results in photogrammetry and view synthesis., Doctorat en Sciences de l'ingénieur et technologie, info:eu-repo/semantics/nonPublished

This paper presents the LPMLEn, a new method to estimate vertical flux and thermal diffusivity from streambed temperature time series using the frequency domain. The main advantages of this new method are: (a) the use of multiple frequencies and multiple sensors for the parameter estimation; (b) noise/uncertainty handling in an optimal way; (c) the possibility to estimate the parameters with both semi-infinite and bounded domain models; and (d) the compensation for temperature drifts in the data known as transients. The capabilities of the LPMLEn are demonstrated using both synthetic and field data, highlighting the advantages of the bounded domain model over the semi-infinite domain model in the parameter estimation process.

The heat flux is one of the key parameter used to quantify and understand transport in fusion devices. In this paper, a new method is introduced to calculate the heat flux including its confidence with high accuracy based on perturbed measurements such as the electron temperature. The new method is based on ideal filtering to optimally reduce the noise contributions on the measurements and piece-wise polynomial approximations to calculate the time derivative. Both methods are necessary to arrive at a heat flux and effective diffusion coefficient with high accuracy. The new methodology is applied to a measurement example using electron cyclotron resonance heating block-wave modulation at the Large Helical Device showing the merit of the newly developed methodology., source:Citation M. van Berkel et al 2018 Nucl. Fusion 58 096036, source:https://doi.org/10.1088/1741-4326/aad13e, identifier:0000-0001-6574-3823

A new methodology to analyze non-linear components in perturbative transport experiments is introduced. The methodology has been experimentally validated in the Large Helical Device for the electron heat transport channel. Electron cyclotron resonance heating with different modulation frequencies by two gyrotrons has been used to directly quantify the amplitude of the non-linear component at the inter-modulation frequencies. The measurements show significant quadratic non-linear contributions and also the absence of cubic and higher order components. The non-linear component is analyzed using the Volterra series, which is the non-linear generalization of transfer functions. This allows us to study the radial distribution of the non-linearity of the plasma and to reconstruct linear profiles where the measurements were not distorted by non-linearities. The reconstructed linear profiles are significantly different from the measured profiles, demonstrating the significant impact that non-linearity can have., source:Citation M. van Berkel et al 2017 Nucl. Fusion 57 126036, source:https://doi.org/10.1088/1741-4326/aa827a, identifier:0000-0001-6574-3823

This paper introduces a recent innovation in dealing with non-periodic behavior often referred to as transients in perturbative experiments. These transients can be the result from the unforced response due to the initial condition and other slow trends in the measurement data and are a source of error when performing and interpreting Fourier spectra. Fourier analysis is particularly relevant in system identification used to build feedback controllers and the analysis of various pulsed experiments such as heat pulse propagation studies. The basic idea behind the methodology is that transients are continuous complex-valued smooth functions in the Fourier domain which can be estimated from the Fourier data. Then, these smooth functions can be subtracted from the data such that only periodic components are retained. The merit of the approach is shown in two experimental examples, i.e. heat pulse propagation (core transport analysis) and radiation front movement due to gas puffing in the divertor. The examples show that the quality of the data is significantly improved such that it allows for new interpretation of the results even for non-ideal measurements., source:M van Berkel et al 2020 Plasma Phys. Control. Fusion 62 094001, source:https://doi.org/10.1088/1361-6587/ab9eaa, identifier:0000-0001-6574-3823