SDM-Car: A Dataset for Small and Dim Moving Vehicles Detection in Satellite Videos

2024-12-24

IEEE Geoscience and Remote Sensing Letters

AI Summary

Weak-radiance, low-contrast, small-scale moving vehicles—especially under dim illumination—suffer severe detection performance degradation in satellite video. To address this, we introduce SDM-Car, the first dedicated benchmark dataset for such targets, comprising 99 high-resolution remote sensing video sequences captured by the Luojia-3 01 satellite and featuring systematic annotation of “small-and-dim” dynamic vehicles. We propose an enhanced YOLOv8 framework integrating multi-scale image enhancement with a channel-spatial joint attention mechanism to jointly improve feature robustness under low signal-to-noise ratio and temporal consistency. SDM-Car is publicly released. Our method achieves a 12.7% mAP gain over baseline YOLOv8 on SDM-Car, outperforming state-of-the-art detectors. Moreover, failure mode analysis reveals critical limitations of existing models in weak-radiance scenarios, thereby bridging key gaps in both data and algorithmic foundations for detecting small, dim moving objects in remote sensing video.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Mobility, Driving & Flight

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements

Abstract

Vehicle detection and tracking in satellite video is essential in remote sensing (RS) applications. However, upon the statistical analysis of existing datasets, we find that the dim vehicles with low radiation intensity and limited contrast against the background are rarely annotated, which leads to the poor effect of existing approaches in detecting moving vehicles under low radiation conditions. In this letter, we address the challenge by building a small and dim moving cars (SDM-Car) dataset with a multitude of annotations for dim vehicles in satellite videos, which is collected by the Luojia 3–01 satellite and comprises 99 high-quality videos. Furthermore, we propose a method based on image enhancement and attention mechanisms to improve the detection accuracy of dim vehicles, serving as a benchmark for evaluating the dataset. Finally, we assess the performance of several representative methods on SDM-Car and present insightful findings. The dataset is openly available at https://github.com/TanedaM/SDM-Car.

Problem

• Vehicle Detection
• Low-Light Conditions
• Satellite Imagery

Innovation

• SDM-Car Dataset
• Enhanced Imaging Technique
• Low-Light Vehicle Detection