An aerial color image anomaly dataset for search missions in complex forested terrain

2025-07-21

AI Summary

Detecting and localizing minute anomalies—such as forensic clues or rescue targets—within dense forest vegetation under complex terrain remains highly challenging due to severe occlusion. Method: This work introduces the first publicly available, high-resolution aerial color image dataset for anomaly detection specifically designed for realistic occluded scenarios. It integrates high-resolution aerial imaging, crowdsourced manual annotation, offline preprocessing, and online interactive labeling, supported by a dynamically extensible web-based interface. Contribution/Results: Extensive experiments reveal significant deficiencies in existing anomaly detection methods regarding contextual awareness. Beyond systematically diagnosing these limitations, this work establishes an authoritative benchmark that advances research on context-aware anomaly detection models tailored for occluded environments. The dataset and interface facilitate reproducible evaluation and iterative model development, thereby bridging a critical gap between real-world deployment needs and current algorithmic capabilities.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Computer Vision: Segmentation
• Humans and AI: Crowd Sourcing and Human Computation

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

After a family murder in rural Germany, authorities failed to locate the suspect in a vast forest despite a massive search. To aid the search, a research aircraft captured high-resolution aerial imagery. Due to dense vegetation obscuring small clues, automated analysis was ineffective, prompting a crowd-search initiative. This effort produced a unique dataset of labeled, hard-to-detect anomalies under occluded, real-world conditions. It can serve as a benchmark for improving anomaly detection approaches in complex forest environments, supporting manhunts and rescue operations. Initial benchmark tests showed existing methods performed poorly, highlighting the need for context-aware approaches. The dataset is openly accessible for offline processing. An additional interactive web interface supports online viewing and dynamic growth by allowing users to annotate and submit new findings.

Problem

• Detecting small anomalies in dense forest aerial imagery
• Improving automated search in occluded, complex terrains
• Benchmarking context-aware anomaly detection for manhunts

Innovation

• High-resolution aerial imagery for dense forests
• Crowd-search initiative for anomaly detection
• Interactive web interface for dynamic annotations