Fraunhofer-Institut für Optronik, Systemtechnik und Bildauswertung IOSB

Filters

4
Reset filters

Settings
Now showing 1 - 10 of 2791
  • Patent
    Network node for a non-detectable laser communication system
    ( 2024-02-07)
    Kieleck, Christelle
    ;
    Eichhorn, Marc
    ;
    Rudow, Oliver
    ;
    Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.
    ;
    Hensoldt Sensors GmbH
    A network node (120) for a non-detectable laser communication system (100), wherein the laser communication system (100) is configured to send to the network node (120) at least one laser beam (10), comprises a reflector device (123), configured to generate, by a reflection of the laser beam (10), a reflected laser beam (20), and a modulator device (125), configured to provide a modulation of the reflected laser beam (20).
  • Publication
    Designing User Interfaces for Automated Driving: A Simulator Study on Individual Information Preferences
    ( 2024)
    Driesen Micklitz, Tim
    ;
    Fellmann, Michael
    ;
    Röcker, Carsten
    Automated Driving (AD) can free users from driving and create time for their disposal. However, since manufacturers increasingly target a wide customer range with AD, such systems and their User Interfaces (UI) must accommodate different user characteristics and preferences. This paper aims to analyze the effects of individual characteristics on the information preferences in UIs for surrounding road infrastructure (for instance, lane markings or traffic signs) and system limits describing hindering factors for AD (for instance, construction sites or unsuitable weather conditions). To do so, we performed a driving simulator study with 43 participants. Results show that users with a more positive attitude towards technology prefer more infrastructure information. Furthermore, users familiar with Automatic Cruise Control prefer less system limit information, while higher experience with Steering Assists relates to higher preference in this regard. These findings add concrete mechanisms to the theory of personalized AD UIs and inform product development on how to create more personalized user experiences. By this, we aim to address challenges regarding the acceptance, adoption, and usage of AD.
  • Publication
    Temperature investigation of low SWaP thulium-doped fiber lasers
    ( 2024)
    Panitzek, Dieter
    ;
    Romano, Clément
    ;
    Eichhorn, Marc
    ;
    Kieleck, Christelle
    We investigate the temperature dependence of an in-band core-pumped thulium-doped fiber laser with a low SWaP (size, weight, and power) architecture. The temperature investigation is carried out both experimentally and numerically by a simulation model. We demonstrate experimentally that the investigated setup is resistant for temperatures till 353 K. In addition, we explain the observed behavior by considering the temperature depended spectroscopic parameters of thulium-doped silica fibers. Finally, a numerical investigation is carried out for higher temperatures up to 573 K and higher output powers up to 12 W as well as for different wavelengths and show that the considered fiber lasers works still efficient at these temperature ranges. We show the reliability of the considered thulium-doped fiber laser architecture for applications in harsh environment.
  • Publication
  • Publication
    Tiled aperture coherent beam combination of 2 μm fiber lasers
    ( 2024)
    Romano, Clément
    ;
    Panitzek, Dieter
    ;
    Pradat-Peyre, Gabriel
    ;
    Milcent, Simon
    ;
    Uthayakumar, Jashaani
    ;
    Eichhorn, Marc
    ;
    Kieleck, Christelle
    Exceeding the multi-kW power level with thulium-doped fiber lasers has not been achieved using a single thulium-doped fiber laser. One solution to overcome this limit is the coherent beam combination. We focus on an active phase control with tiled aperture configuration. The setup consists in an amplified seed laser split in three channels. These channels are controlled in phase and amplified again before being launched free space and combined. A SPGD algorithm controls the channel’s phase to provide combination. Rise time below 0.5 ms were achieved with a residual amplitude noise lower than λ/30.
  • Publication
    Knowledge-Distillation-Based Label Smoothing for Fine-Grained Open-Set Vehicle Recognition
    ( 2024)
    Wolf, Stefan
    ;
    Loran, Dennis
    ;
    Beyerer, Jürgen
    Fine-grained vehicle classification describes the task of estimating the make and the model of a vehicle based on an image. It provides a useful tool for security authorities to find suspects in surveillance cameras. However, most research about fine-grained vehicle classification is only focused on a closed-set scenario which considers all possible classes to be included in the training. This is not realistic for real-world surveillance applications where the images fed into the classifier can be of arbitrary vehicle models and the large number of commercially available vehicle models renders learning all models impossible. Thus, we investigate fine-grained vehicle classification in an open-set recognition scenario which includes unknown vehicle models in the test set and expects these samples to be rejected. Our experiments highlight the importance of label smoothing for open-set recognition performance. Nonetheless, it lacks recognizing the different semantic distances between vehicle models which result in largely different confusion probabilities. Thus, we propose a knowledge-distillation-based label smoothing approach which considers these different semantic similarities and thus, improves the closed-set classification as well as the open-set recognition performance.
  • Publication
    Professionelle Leitung von KI-Projekten in der Industrie - wie geht Systems Engineering mit KI?
    ( 2024)
    Pfrommer, Julius
    ;
    Usländer, Thomas
    Inspiriert von den Erfolgsmeldungen jüngerer Zeit entschließen sich immer mehr Unternehmen selbst KI-Initiativen zu starten. An vielen Stellen kann die KI ihren Wert unter Beweis stellen - und doch ist es eine Herausforderung über einen Proof-of-Concept hinaus bis in den langfristigen Betrieb zu kommen. An dieser Stelle sind insbesondere die Projektleitenden gefragt Anforderungen zu antizipieren und abzustimmen, Datenbedarfe und Aufwände zu schätzen und die Verknüpfung KI-basierter Entwicklung mit klassischen Ingenieurdisziplinen zu koordinieren. Hierfür gibt es eine Reihe von Werkzeugen und Frameworks, die die Projektleitenden in dieser Aufgabe unterstützen.
  • Publication
    Pattern Recognition. Introduction, Features, Classifiers and Principles
    (De Gruyter, 2024)
    Beyerer, Jürgen
    ;
    Hagmanns, Raphael
    ;
    Stadler, Daniel
    The book offers a thorough introduction to Pattern Recognition aimed at master and advanced bachelor students of engineering and the natural sciences. Besides classification - the heart of Pattern Recognition - special emphasis is put on features: their typology, their properties and their systematic construction. Additionally, general principles that govern Pattern Recognition are illustrated and explained in a comprehensible way. Rather than presenting a complete overview over the rapidly evolving field, the book clarifies the concepts so that the reader can easily understand the underlying ideas and the rationale behind the methods. For this purpose, the mathematical treatment of Pattern Recognition is pushed so far that the mechanisms of action become clear and visible, but not farther. Therefore, not all derivations are driven into the last mathematical detail, as a mathematician would expect it. Ideas of proofs are presented instead of complete proofs. From the authors’ point of view, this concept allows to teach the essential ideas of Pattern Recognition with sufficient depth within a relatively lean book.
  • Publication
    Investigation of the polymer material perforation time: comparison between two fiber laser wavelengths
    ( 2024)
    Romano, Clément
    ;
    Ritt, Gunnar
    ;
    Henrichsen, Michael
    ;
    Eichhorn, Marc
    ;
    Kieleck, Christelle
    This study investigated the perforation time of polyamide 6.6 using fiber lasers at two different wavelengths: 1070 and 1943 nm. The novelty of this research lies in the comparison of perforation times at equivalent laser irradiances on the polymer sample with two different colors of polyamide 6.6: natural and black. The results revealed that, at comparable irradiance levels and beam diameters, the 1943 nm laser source perforated the polyamide 6.6 sample faster than the 1070 nm laser source. The difference in perforation time was found to be significantly higher for natural-colored polyamide 6.6 compared to black-colored polyamide 6.6. These findings suggest that, for material processing of polyamide 6.6, especially in terms of perforation, the use of 2 μm laser sources should be privileged over 1 μm laser sources.
  • Publication
    An end-to-end machine learning approach with explanation for time series with varying lengths
    ( 2024)
    Schneider, Manuel
    ;
    Greifzu, Norbert
    ;
    Wang, Lei
    ;
    Walther, Christian
    ;
    Wenzel, Andreas
    ;
    Li, Pu
    An accurate prediction of complex product quality parameters from process time series by an end-to-end learning approach remains a significant challenge in machine learning. A special difficulty is the application of industrial batch process data because many batch processes generate variable length time series. In the industrial application of such methods, explainability is often desired. In this study, a 1D convolutional neural network (CNN) algorithm with a masking layer is proposed to solve the problem for time series of variable length. In addition, a novel combination of 1D CNN and class activation mapping (CAM) technique is part of this study to better understand the model results and highlight some regions of interest in the time series. As a comparative state-of-the-art unsupervised machine learning method, the One-Nearest Neighbours (1NN) algorithm combined with dynamic time warping (DTW) was used. Both methods are investigated as end-to-end learning methods with balanced and unbalanced class distributions and with scaled and unscaled input data, respectively. The FastDTW and DTAIDistance algorithms were investigated for the DTW calculation. The data set is made up of sensor signals that was collected during the production of plastic parts. The objective was to predict a quality parameter of plastic parts during production. For this research, the quality parameter will be a difficult or only destructively measurable parameter and both methods will be investigated for their applicability to this prediction task. The application of the proposed approach to an industrial facility for producing plastic products shows a prediction accuracy of 83.7%. It can improve the reverence method by approximately 1.4%. In addition to the slight increase in accuracy, the CNN training time was significantly reduced compared to the DTW calculation.