Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Wouter Jansen has a total of 50 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Synthetic Data for Data-Efficient Agricultural Computer Vision: Three Datasets and a Multi-Task Benchmark Across Real and Simulated Domains

    by Steven Moonen, Wouter Jansen, Wenzhi Liao, Mohammad Hasan Rahmani, Ayyoub Ahar, Abdellatif Bey-Temsamani, Jan Steckel, Nick Michiels

    The availability of large annotated datasets remains a major bottleneck for deploying computer vision systems in agricultural and agri-food applications, particularly for tasks requiring fine-grained annotations such as defect detection and instance segmentation. Synthetic data generation offers a scalable alternative, but its effectiveness in complex real-world scenarios remains insufficiently understood. In this work, the role of synthetic data is investigated across three representative tasks each posing unique challenges: apple detection in orchard environments, potato–stone classification in industrial sorting, and carrot crack detection for quality inspection. The real and synthetic data, as well as their annotations, are publicly available. The synthetic data is created using a set of tools for fast large-scale procedural scene generation with highly detailed natural assets. For each task, datasets are constructed combining real and synthetically generated images with pixel-level annotations. Training strategies are systematically evaluated using fully real, limited real, synthetic-only, and combined datasets. The results show that synthetic-only training leads to a clear performance gap on real-world data, with mAP50 decreasing from 0.866–0.891 for fully real training to 0.396–0.703 for synthetic-only training across the three use cases. However, combining synthetic data with only 10 real images recovers 75–98% of the gap between training on those 10 real images alone (mAP50 0.194–0.497) and full real training, reaching mAP50 values of 0.723–0.885. These findings highlight the potential of synthetic data for improving data efficiency and provide a reproducible benchmark for future research on sim-to-real transfer in agricultural computer vision.


  • Towards semantic acoustic perception : simulation, embedded processing, and robotic applications of in-air 3D sonar

    by Wouter Jansen

    The pursuit of autonomous mobile robotics demands perception systems capable of operating in harsh, optically denied environments. While optical sensors such as cameras and Lidar provide exceptionally high spatial resolution, they suffer severe performance degradation in the presence of dust, fog, or smoke. Because ultrasonic acoustic signals have much longer wavelengths, they are a highly resilient alternative. Despite this, the adoption of 3D Sonar remains limited. We believe this is partially due to the modality's lack of semantic interpretability. To enable context-aware navigation, foundational steps must be taken to elevate acoustic perception from a purely spatial tool into a semantic modality capable of interpreting complex environments. This dissertation presents a comprehensive framework facilitating these steps toward robust, context-aware acoustic perception. The contributions are structured around three pillars. Firstly, beyond standard synthetic data generation, we introduce a high-fidelity multi-modal simulation framework. Incorporating a hardware-accelerated acoustic ray tracing model, this framework enables closed-loop testing of autonomous systems within dynamic, high-resolution environments. By accurately modeling complex surface material properties and reflectance for both optical and acoustic sensors, it facilitates rigorous cross-modal algorithm validation and specialized ensonification experiments. Secondly, a GPU-accelerated real-time 3D Sonar platform is developed for harsh conditions. Its specialized architecture achieves the high spatial resolution and signal-to-noise ratio necessary for active perception within strict real-time constraints. Finally, this research explores multi-modal integration, distributed sensing, and acoustic feature extraction as steps toward the development of semantic autonomous systems. To enhance spatial awareness and mobility, a real-time multi-Sonar navigation algorithm utilizing acoustic flow queues enables decentralized sensor placement. Furthermore, fusing Sonar data with an inertial measurement unit achieves acoustic image stabilization, maintaining reliable perception on uneven terrain. Moving toward semantic recognition, this work explores the classification of bio-inspired artificial acoustic landmarks and reflector constellations, ultrasonic counterparts to QR codes. By training convolutional neural networks on cochleogram representations of raw acoustic echoes, these engineered objects establish a baseline for extracting structural information from broadband signals. Collectively, these contributions lay the groundwork for the deployment of reliable, context-aware mobile robots in unstructured environments with acoustic sensors.

  • Open Access

    Hardware-Accelerated Geometrical Simulation of Biological and Engineered In-Air Ultrasonic Systems

    by Wouter Jansen, Jan Steckel

    The deployment of in-air acoustic sensors for industrial monitoring and autonomous robotics has grown significantly, often drawing inspiration from biological echolocation. However, developing and validating these systems in existing simulation frameworks remains challenging due to the computational cost of simulating high-frequency wave propagation in large, dynamic, and complex environments. While wave-based methods offer high accuracy, they scale poorly with frequency and volume. Conversely, existing geometric acoustic solvers often lack support for dynamic scenes, complex diffraction, or closed-loop robotic integration. In this work, we introduce SonoTraceUE, a high-fidelity acoustic simulation framework built as a plugin for Unreal Engine. By using a hardware-accelerated ray tracing-based specular reflection model, and a curvature-based Monte Carlo diffraction model, the system enables near real-time simulation of active and passive acoustic sensing in dynamic, multi-material environments. We validate the framework through two distinct experimental domains: a bioacoustic study and a robotics experiment. Our results demonstrate that SonoTraceUE achieves high correlation with real-world spectral and spatial data. The framework provides a versatile platform for synthetic data generation, hypothesis testing in bioacoustics, and the rapid prototyping of closed-loop robotic systems that use acoustic sensing.

  • Open Access

    LLM-Driven 3D Scene Generation of Agricultural Simulation Environments

    by Arafa Yoncalik, Wouter Jansen, Nico Huebel, Mohammad Hasan Rahmani, Jan Steckel

    Procedural generation techniques in 3D rendering engines have revolutionized the creation of complex environments, reducing reliance on manual design. Recent approaches using Large Language Models (LLMs) for 3D scene generation show promise but often lack domain-specific reasoning, verification mechanisms, and modular design. These limitations lead to reduced control and poor scalability. This paper investigates the use of LLMs to generate agricultural synthetic simulation environments from natural language prompts, specifically to address the limitations of lacking domain-specific reasoning, verification mechanisms, and modular design. A modular multi-LLM pipeline was developed, integrating 3D asset retrieval, domain knowledge injection, and code generation for the Unreal rendering engine using its API. This results in a 3D environment with realistic planting layouts and environmental context, all based on the input prompt and the domain knowledge. To enhance accuracy and scalability, the system employs a hybrid strategy combining LLM optimization techniques such as few-shot prompting, Retrieval-Augmented Generation (RAG), finetuning, and validation. Unlike monolithic models, the modular architecture enables structured data handling, intermediate verification, and flexible expansion. The system was evaluated using structured prompts and semantic accuracy metrics. A user study assessed realism and familiarity against real-world images, while an expert comparison demonstrated significant time savings over manual scene design. The results confirm the effectiveness of multi-LLM pipelines in automating domain-specific 3D scene generation with improved reliability and precision. Future work will explore expanding the asset hierarchy, incorporating real-time generation, and adapting the pipeline to other simulation domains beyond agriculture.

  • Open Access

    Delay-Multiply-and-Sum Beamforming for Real-Time In-Air Acoustic Imaging

    by Wouter Jansen, Walter Daems, Jan Steckel

    In-air acoustic imaging systems demand beamforming techniques that offer high dynamic range and spatial resolution while remaining robust, yet conventional Delay-and-Sum (DAS) beamforming suffers from high sidelobes and low contrast. Conversely, advanced adaptive methods are typically too computationally expensive for real-time field operation. To overcome this trade-off, we propose a higher-order non-linear beamforming method using the Delay-Multiply-and-Sum (DMAS) technique with Coherence Factor weighting, specifically adapted for in-air microphone arrays. By leveraging Newton-Girard binomial expansion, we reduce the complexity of n-th order DMAS to O(N), enabling GPU-accelerated, real-time performance on embedded platforms, with the implementation also made available as open-source. Through validation against simulated and real-world data, we demonstrate that the proposed 5th-order DMAS with CF achieves a dynamic range improvement of nearly 40 dB over DAS, alongside significantly higher image SNR and artifact suppression. We demonstrate real-time viability on various GPU platforms, establishing this approach as a practical, high-performance solution for 3D in-air sonar and industrial acoustic imaging.


  • AirleakSIM: Using Cosys-AirSim to Benchmark Different Pressurized Airleak Localization Strategies

    by Anthony Schenck, Wouter Jansen, Walter Daems, Jan Steckel

    Leaks in pressurized air systems cause substantial energy losses in industrial environments, with detection typically relying on labor-intensive manual inspection methods. To address this, autonomous systems for leak localization have been proposed, though their performance is often difficult to evaluate due to the lack of real-world ground truth. This paper introduces AirleakSIM, a simulation environment built upon Cosys-AirSim, an Unreal Engine-based platform extended with realistic acoustic ray tracing and sensor modeling, used to benchmark ultrasonic sound source localization algorithms. We simulate pressurized air leaks in a large-scale industrial environment, including complex acoustic propagation effects, and generate realistic microphone array data. This data is used to evaluate the performance of four different beamforming algorithms (delay-and-sum, MUSIC, MVDR and CLEAN) across twenty six levels of noise and for two different microphone array geometries. Our results show that advanced beamformers such as MUSIC and MVDR surpass traditional methods in noisy conditions, and that a random microphone array topology outperforms a microphone array using a rectangular grid topology.


  • Ruggedized Embedded Ultrasound Sensing in Harsh Conditions: eRTIS in the Wild

    by Dennis Laurijssen, Wouter Jansen, Arne Aerts, Walter Daems, Jan Steckel

    We present eRTIS, a rugged, embedded ultrasound sensing system for use in harsh industrial environments. The system features a broadband capacitive transducer and a 32-element MEMS microphone array capable of 2D and 3D beamforming. A modular hardware architecture separates sensing and processing tasks: a high-performance microcontroller handles excitation signal generation and data acquisition, while an NVIDIA Jetson module performs GPU-accelerated signal processing. eRTIS supports external synchronization via a custom controller that powers and coordinates up to six devices, either simultaneously or in a defined sequence. Additional synchronization options include bidirectional triggering and in-band signal injection. A sealed, anodized aluminum enclosure with passive cooling and IP-rated connectors ensures reliability in challenging conditions. Performance is demonstrated in three field scenarios: harbor mooring, off-road robotics, and autonomous navigation in cluttered environments, demonstrates that eRTIS provides robust sensing in situations where optical systems degrade.


  • Open Access

    Towards In-Air Ultrasonic QR Codes: Deep Learning for Classification of Passive Reflector Constellations

    by Wouter Jansen, Jan Steckel

    In environments where visual sensors falter, in-air sonar provides a reliable alternative for autonomous systems. While previous research has successfully classified individual acoustic landmarks, this paper takes a step towards increasing information capacity by introducing reflector constellations as encoded tags. Our primary contribution is a multi-label Convolutional Neural Network (CNN) designed to simultaneously identify multiple, closely spaced reflectors from a single in-air 3D sonar measurement. Our initial findings on a small dataset confirm the feasibility of this approach, validating the ability to decode these complex acoustic patterns. Secondly, we investigated using adaptive beamforming with null-steering to isolate individual reflectors for single-label classification. Finally, we discuss the experimental results and limitations, offering key insights and future directions for developing acoustic landmark systems with significantly increased information entropy and their accurate and robust detection and classification.

  • FL-RTIS, a Novel Multimodal Sensor Using High-Speed Camera and Active 3-D Sonar for Insect Ensonification

    by Jan Steckel, Pamela Rivera Parra, Arne Aerts, Dennis Laurijssen, Wouter Jansen, Walter Daems, Jesse Barber

    In this letter, we introduce the flutter real-time imaging sonar (FL-RTIS): a novel sensor system that integrates a high-speed camera with a 3-D sonar sensor to investigate insect ensonification. By capturing and synchronizing high-resolution video with dense 3-D acoustic data, FL-RTIS provides a detailed analysis of the echo dynamics from fluttering insects. This multimodal approach allows for an unprecedented study of the acoustic interactions between bats and their prey, facilitating more profound insights into evolutionary adaptations in predator-prey dynamics. The capabilities of the FL-RTIS are demonstrated through laboratory experiments and field tests, highlighting its potential for gathering large datasets and showing the potential for new avenues to understanding complex biological interactions.