Filters

Settings

Now showing1 - 10 of 14
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
    SynthDefekt: A Gen-AI-Assisted Pipeline for Generating Synthetic Data for Industrial Visual Inspection
    Supervised learning methods are the de facto standard in industrial machine vision, despite challenging boundary conditions (varying/uncontrolled illumination, moving objects, unknown viewpoints). Their performance, however, depends critically on training data that are of sufficient quality, adequately diverse, and correctly annotated. In visual inspection tasks, defect images are often not available in sufficient quantity and variety: rare cases are underrepresented. Machine vision systems must reliably detect these rare cases as well. Synthetic, computer-generated data are a promising alternative: rare defect classes can be generated in a targeted manner, annotations are available “by design,” application-relevant variations can be covered systematically, generation is scalable and IP-compliant, and it enables controlled class balancing at high variance. A common approach is image synthesis via computer graphics. In rule-based, 3D scene–based synthesis, however, extensive manual modeling effort is required (scene geometry/composition, material and texture models, defect models), and thus expertise spanning computer graphics, computer vision, and domain knowledge. In prior work, we were able to show that domain-specifically generated datasets enable specialist models to perform robustly in real-world visual inspection. SynthDefekt addresses the high manual effort through a semi-automated, coupled pipeline: procedural scene composition, object modification, defect generation, and material modeling are linked with post-processing (image enhancement, augmentation, domain adaptation), and the suitability of the generated dataset is quantified. The end-to-end process chain subsequently allows optimization. In an iterative workflow, besides rendering time, evaluation effort as well as time and compute resources are critical. The gold standard is to train ML models on the generated image data and validate them on real data in order to quantify suitability directly. To accelerate the evaluation step and enable more optimization cycles, we investigate to what extent suitability can be assessed indirectly with sufficient accuracy—for example by fusing or sequentially applying similarity metrics (FID, KID), domain-adaptation metrics (contrastive domain loss), semantic consistency measures (CLIP embedding distances), and proxy tasks. The development of this evaluation module constitutes the second project objective. In addition, GenAI assistance is being explored for manual sub-steps, such as pipeline configuration. Depending on technological maturity, diffusion models (e.g., for generating material models) or LLMs (e.g., for script synthesis for scene composition) are to be employed, among other methods.
      7
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      10
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      28
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      16
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      16
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
    Load-oriented additive manufacturing of optimized fiber-reinforced composites
    Due to the high achievable stiffness and strength at low weight, carbon fiber reinforced plastic (CFRP) is an attractive material for lightweight construction. Classical manufacturing processes such as hand lay-up, winding or tape laying are often limited in their applicability for complex geometric structures, such as those resulting from structural optimization. Additive manufacturing of CFRP using fused deposition modeling offers the possibility of placing individual fiber strands in the load direction to increase the effectiveness of material use. Furthermore, additive manufacturing enables accelerated development processes and the economical production of small series. The aim of the research project is to overcome this limitation through non-planar printing on machines with higher degrees of freedom. To this end, a manufacturing process for the additive manufacturing of CFRP components is to be developed, with which volumetric bodies can be produced from non-planar layers in a highly automated manner for the first time, taking load case analyses into account. The focus is on the slicing and path planning software and the necessary investigation of the manufacturing process.
      10
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
    Integrated structural optimization and model-informed path planning of additively manufactured, continuous fiber-reinforced plastics, taking into account manufacturing restrictions and design-dependent strength
    For structures with complex fiber architecture, the transfer of a fiber composite design into the path planning of an additive manufacturing process in particular has been a mostly manual process to date. For example, the result of optimizing a fiber composite structure is available in the form of a vector field of local material orientations. Manual path planning is then used to implement a component that comes close to this design. Today, there is no approach for optimizing additively manufactured composite components made of continuous fibers in such a way that a manufacturable design is created and at the same time its full lightweight potential is exploited. The weak point here is the transition between the structural-mechanical model, path planning and the manufacturing process, which is why a continuous design chain can only be developed in close cooperation between structural optimization and manufacturing technology. The overarching aim of this application is therefore to develop a continuous process for model-informed path planning and structural optimization of additively manufactured fibre composite components. The path planning should anticipate and compensate for process-induced sources of error through simulation, control and modeling of the non-planar production and the fiber-matrix combination. In addition, the design-dependent, local strength and manufacturing restrictions are to be taken into account in the optimization. To this end, a method for error-compensating path planning with model-based process parameters is being developed and a fused filament fabrication process for the non-planar additive manufacturing of fiber-reinforced plastic composites is adapted to it. The paths determined in this way are the input for a method to be developed for the structural optimization of fiber paths while maintaining manufacturability and taking into account design-dependent strengths and stiffnesses. This overall process is to be validated by production trials and structural-mechanical tests.
      10
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      7
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
    DrumView
    Gemeinsam arbeiten die Projektpartner an einer möglichst automatisierten und vollständigen Digitalisierung der Flugzeugkabine, d. h. der Aufnahme vieler einzelner Fotos, die anschließend in einem Modell zusammengeführt werden. Hierzu gilt es geeignete Kameras und weitere benötigte Sensoren auszuwählen. Um eine hohe Aufnahmequalität zu erreichen wird eine Kinematik entwickelt, welche die systematische Führung der Sensorik durch die Flugzeugkabine gestattet. So können basierend auf den hochwertigen Daten KI-Modelle trainiert werden, die die Detektion von Anomalien erlauben. Mögliche Einsatzszenarien des Systems liegen bspw. bei den Flugzeugherstellern, die die Kabine im Montageprozess oder insbesondere vor der Übergabe an den Kunden inspizieren wollen, um etwaige Schäden frühzeitig zu erkennen und eine vollständige und nachvollziehbare Dokumentation abzuleiten. Zu möglichen Schäden, die zur Beanstandung durch den Kunden führen können, gehören bspw. Verschmutzungen, Kratzer auf Oberflächen, Spaltmaße oder falsche montierte Komponenten.
      10
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Project, Access status: without files ,
      7