FOD-S2R: A FOD Dataset for Sim2Real Transfer Learning based Object Detection

TL;DR

This work introduces FOD-S2R, a hybrid dataset with real and Unreal Engine–generated synthetic images to study foreign object debris detection inside aircraft fuel tanks. It benchmarks a range of anchor-based and anchor-free detectors, analyzes the domain gap between synthetic and real data, and demonstrates that synthetic data are most effective as pretraining for real-world performance. Sim2Real transfer experiments show a notable improvement in real-world detection accuracy when synthetic data are used for pretraining, while reversing the order can degrade learned representations. The dataset and findings offer a scalable path toward robust, automated FOD inspection in enclosed aviation environments, reducing reliance on extensive real-data collection.

Abstract

Foreign Object Debris (FOD) within aircraft fuel tanks presents critical safety hazards including fuel contamination, system malfunctions, and increased maintenance costs. Despite the severity of these risks, there is a notable lack of dedicated datasets for the complex, enclosed environments found inside fuel tanks. To bridge this gap, we present a novel dataset, FOD-S2R, composed of real and synthetic images of the FOD within a simulated aircraft fuel tank. Unlike existing datasets that focus on external or open-air environments, our dataset is the first to systematically evaluate the effectiveness of synthetic data in enhancing the real-world FOD detection performance in confined, closed structures. The real-world subset consists of 3,114 high-resolution HD images captured in a controlled fuel tank replica, while the synthetic subset includes 3,137 images generated using Unreal Engine. The dataset is composed of various Field of views (FOV), object distances, lighting conditions, color, and object size. Prior research has demonstrated that synthetic data can reduce reliance on extensive real-world annotations and improve the generalizability of vision models. Thus, we benchmark several state-of-the-art object detection models and demonstrate that introducing synthetic data improves the detection accuracy and generalization to real-world conditions. These experiments demonstrate the effectiveness of synthetic data in enhancing the model performance and narrowing the Sim2Real gap, providing a valuable foundation for developing automated FOD detection systems for aviation maintenance.

Figures (11)

• Figure 1: Unreal Engine editor used to assemble fuel-tank scenes for synthetic FOD data generation, allowing precise control over camera pose, lighting, and object layout.
• Figure 2: Blueprint scripting for Dataset generation using Unreal Engine.
• Figure 3: The synthetic images in FOD-S2R showcase a range of lighting conditions, diverse glove colors, and variations in the color of the fuel tank, as seen in the image.
• Figure 4: The real dataset with varied lightning conditions from different angles. The real dataset includes diverse FOV's, FOD size, lightening conditions etc.
• Figure 5: Instances per category.