Fakultät 09 / Institut für Produktentwicklung und Konstruktionstechnik
Refine
Language
- English (16)
Has Fulltext
- yes (16)
Keywords
- 3D Printing (2)
- 3D Printed Electronics (1)
- 3D-Druck (1)
- 60J10 (1)
- 60K30 (1)
- Adaption (1)
- Adaptive Thresholding (1)
- Albedo (1)
- Anthropomorphic robotic hands (1)
- Artificial Intelligence (1)
Faculty
- Fakultät 09 / Institut für Produktentwicklung und Konstruktionstechnik (16)
- Fakultät 03 / Institut für Informationswissenschaft (1)
- Fakultät 09 / Cologne Institute for Renewable Energy (1)
- Fakultät 09 / Institut für Rettungsingenieurwesen und Gefahrenabwehr (1)
- Fakultät 09 / Institut für Werkstoffanwendung (1)
This paper presents a comparative study of binarization techniques for automated defect detection in dye penetrant testing (DPT) images. We evaluate established methods, including global, adaptive, and histogram-based thresholding, against three novel machine learning-assisted approaches, Soft Binarization (SoBin), Delta Binarization (DeBin), and Convolutional Autoencoder Binarization (AutoBin), using a real-world dataset from an automated DPT system inspecting stainless steel pipes. Performance is assessed with both pixel-level and region-level metrics, with particular emphasis on the influence of defect saturation. Defect saturation is quantified as the mean saturation value of all pixels belonging to a given defect, and defects are grouped into ten categories spanning from low (60–68) to high (132–140) mean saturation. Our results demonstrate that for lower mean defect saturation values, methods such as AutoBin_Triangle , HSV_global_70 , and SoBin achieve superior Intersection over Union (IoU) and high true positive rates. In contrast, methods based primarily on global thresholding of the saturation channel tend to perform competitively on images with higher defect saturation levels, reflecting their sensitivity to stronger color signals. Moreover, depending on the method, nearly perfect region-level true positive rates ( TPRregion) or minimal false positive rates ( FPRregion) can be attained, emphasizing the trade-off that different models offer distinct strengths and weaknesses, which necessitates selecting the optimal method based on the specific quality control requirements and risk tolerances of the industrial process. These findings underscore the critical importance of defect saturation as a cue for both human and computer vision systems and provide valuable insights for developing robust automated quality control and predictive quality algorithms.
As global reliance on sustainable energy solutions intensifies, there is a growing need to optimise and accurately predict renewable energy outputs. Bifacial photovoltaic systems, which are capable of capturing irradiance on both their front and rear sides, represent a significant advancement over traditional monofacial systems, yielding higher energy per area. The accuracy of simulation models for these systems has a direct impact on their financial viability, necessitating the use of comprehensive and reliable simulation frameworks. This research validates BifacialSimu, an open-source simulation tool designed to enhance the prediction of bifacial PV system energy outputs by incorporating multiple simulation models. The practical validation of BifacialSimu is based on empirical data from three diverse geographic locations. The locations of Golden, United States; Heggelbach, Germany; and Florianópolis, Brazil, provide insights into the performance of bifacial PV systems across a range of environmental conditions and installation configurations. These findings underscore the practical applicability of BifacialSimu, with recommendations for simulation model selection and methodological advancements, paving the way for more precise and efficient bifacial PV system simulations across diverse scenarios. This study employs a number of validation metrics, including relative error, coefficient of determination and Normalized Root Mean Square Error, to assess the accuracy of the simulations. The findings indicate that the Ray tracing method is the most accurate of the irradiance simulation modes for most scenarios. The validation results highlight that the Ray Tracing method achieves superior accuracy in irradiance simulations, particularly under varied environmental conditions, while Variable Albedo models further enhance predictive precision by accounting for dynamic factors such as snow cover.
The article focuses explicitly on transformation and analyzes how it can be measured both quantitatively and qualitatively in a case study region in western Germany. It addresses blue, critical, and green infrastructures in a region that was affected by the 2021 floods in Europe. Together with regional actors, co-creative solutions for developing adaptation strategies and infrastructure planning will be developed. Using and combining different conceptual models and applying them to the project region as well as to human infrastructure highlights the different types of change and transformation. This also shows the complexity of such an overall assessment, which needs to include a lot of diverse actors and disciplines. The project's first results are overviews of national and cross-regional changes to infrastructures and administrative structures after the 2021 floods in Germany and at the district level. The interim results reveal that more needs and plans exist than real, measurable transformations and that certain transformations stem from planning long before the floods in 2021. Maps and land use potentials are presented that provide overviews of regional differences in flood, drought, and wildfire exposure and capacities for nature-based solutions. Both the conceptual models developed in this study as well as the application examples could be useful for other researchers and administrative bodies to measure transformation to climate change and other stimuli.
In this paper, we detail the technical development of a conversation design that is sensitive to group dynamics and adaptable, taking into account the subtleties of linguistic variations between dyadic (i.e., one human and one agent) and group interactions in human–robot interaction (HRI) using the German language as a case study. The paper details the implementation of robust person and group detection with YOLOv5m and the expansion of knowledge databases using large language models (LLMs) to create adaptive multi-party interactions (MPIs) (i.e., group–robot interactions (GRIs)). We describe the use of LLMs to generate training data for socially interactive agents including social robots, as well as a self-developed synthesis tool, knowledge expander, to accurately map the diverse needs of different users in public spaces. We also outline the integration of a LLM as a fallback for open-ended questions not covered by our knowledge database, ensuring it can effectively respond to both individuals and groups within the MPI framework.
Vertical indoor farming (VIF) with hydroponics offers a promising perspective for sustainable food production. Intelligent control of VIF system components plays a key role in reducing operating costs and increasing crop yields. Modern machine vision (MV) systems use deep learning (DL) in combination with camera systems for various tasks in agriculture, such as disease and nutrient deficiency detection, and flower and fruit identification and classification for pollination and harvesting. This study presents the applicability of MV technology with DL modelling to detect the growth stages of chilli plants using YOLOv8 networks. The influence of different bird’s-eye view and side view datasets and different YOLOv8 architectures was analysed. To generate the image data for training and testing the YOLO models, chilli plants were grown in a hydroponic environment and imaged throughout their life cycle using four camera systems. The growth stages were divided into growing, flowering, and fruiting classes. All the trained YOLOv8 models showed reliable identification of growth stages with high accuracy. The results indicate that models trained with data from both views show better generalisation. YOLO’s middle architecture achieved the best performance.
Electromagnetic systems, in particular microwave/terahertz sensing technologies, are the newest among nondestructive sensing technologies. Currently, increased attention is pointed towards their use in various applications. Among these, food inspection stands out as a primary area due to its potential risk to human safety. As a result, substantial efforts are currently focused on utilizing microwave/terahertz imaging as a tool to enhance the efficacy of food quality assessments. This paper deals with the exploitation of microwave/terahertz imaging technology for food quality control and assessment. In particular, the work aims at reviewing the latest developments regarding the detection of internal quality parameters, such as foreign bodies, i.e., plastic, glass, and wood substances/fragments, as well as checking the completeness of the packaged food under consideration. Emphasis is placed on the (inline) inspection of wrapped/packaged food, such as chocolates, cookies, pastries, cakes, and similar confectionery products, moving along production conveyor belts. Moreover, the paper gives a recent overview of system prototypes and industrial products and highlights emerging research topics and future application directions in this area.
The importance of radar-based human activity recognition has increased significantly over the last two decades in safety and smart surveillance applications due to its superiority in vision-based sensing in the presence of poor environmental conditions like low illumination, increased radiative heat, occlusion, and fog. Increased public sensitivity to privacy protection and the progress of cost-effective manufacturing have led to higher acceptance and distribution of this technology. Deep learning approaches have proven that manual feature extraction that relies heavily on process knowledge can be avoided due to its hierarchical, non-descriptive nature. On the other hand, ML techniques based on manual feature extraction provide a robust, yet empirical-based approach, where the computational effort is comparatively low. This review outlines the basics of classical ML- and DL-based human activity recognition and its advances, taking the recent progress in both categories into account. For every category, state-of-the-art methods are introduced, briefly explained, and their related works summarized. A comparative study is performed to evaluate the performance and computational effort based on a benchmarking dataset to provide a common basis for the assessment of the techniques’ degrees of suitability.
A novel approach to manufacture components with integrated conductor paths involves embedding and sintering an isotropic conductive adhesive (ICA) during fused filament fabrication (FFF). However, the molten plastic is deposited directly onto the adhesive path which causes an inhomogeneous displacement of the uncured ICA. This paper presents a 3D printing strategy to achieve a homogeneous cross-section of the conductor path. The approach involves embedding the ICA into a printed groove and sealing it with a wide extruded plastic strand. Three parameter studies are conducted to obtain a consistent cavity for uniform formation of the ICA path. Specimens made of polylactic acid (PLA) with embedded ICA paths are printed and evaluated. The optimal parameters include a groove printed with a layer height of 0.1 mm, depth of 0.4 mm, and sealed with a PLA strand of 700 µm diameter. This resulted in a conductor path with a homogeneous cross-section, measuring 660 µm ± 22 µm in width (relative standard deviation: 3.3%) and a cross-sectional area of 0.108 mm2 ± 0.008 mm2 (relative standard deviation 7.2%). This is the first study to demonstrate the successful implementation of a printing strategy for embedding conductive traces with a homogeneous cross-sectional area in FFF 3D printing.
Das Labor für Konstruktionstechnik der Technischen Hochschule Köln beschäftigt sich mit der Mikrodosierung von präzisen Strukturen von Silberleitklebstoff. Hauptsächlich wird bei derartigen Ventilen in berührende und berührungslose Dosierventile unterschieden. Zur Erzeugung dreidimensionaler elektrisch leitfähiger Strukturen können diese Dosierventile an entsprechende Verfahreinheiten adaptiert werden.
We consider a risk model in discrete time with dividends and capital injections. The goal is to maximise the value of a dividend strategy. We show that the optimal strategy is of barrier type. That is, all capital above a certain threshold is paid as dividend. A second problem adds tax to the dividends but an injection leads to an exemption from tax. We show that the value function fulfils a Bellman equation. As a special case, we consider the case of premia of size one. In this case we show that the optimal strategy is a two barrier strategy. That is, there is a barrier if a next dividend of size one can be paid without tax and a barrier if the next dividend of size one will be taxed. In both models, we illustrate the findings by de Finetti’s example.