Exploring Lightweight Structures for Tiny Object Detection in Remote Sensing Images

Abstract

Detecting tiny objects in remote sensing images has been an intriguing yet challenging topic in remote sensing image processing. While significant progress has been made in many studies, most existing methods focus on improving the accuracy of tiny object detection without particular consideration for computational complexity, which restricts their applicability in resource-limited conditions. Therefore, this article aims to design a lightweight detection algorithm tailored for tiny objects in remote sensing images. First, we investigate the impact of the complexity of different components in deep learning-based object detection models on the accuracy of tiny object detection, including the backbone and detection head. Then, a dedicated backbone for tiny object detection is proposed, achieving competitive detection accuracy while remaining lightweight. Moreover, we propose a lightweight detection head that incorporates deformable convolution and optimize the channel dimension. Finally, we combine the above methods to introduce a lightweight network, lightweight tiny object detecection network (LTDNet), for tiny object detection in remote sensing images. Benefiting from the dedicated designs for the backbone and detection head specifically for tiny objects, the proposed method can achieve competitive detection accuracy with very low parameters and computational complexity. Extensive experiments are conducted on the artificial intelligence (AI)-TODv2 and learning vision and remote sensing (LEVIR)-Ship datasets, and the results demonstrate the effectiveness of our proposed method. Specifically, the proposed method achieves 54.6% AP50 on the AI-TODv2 dataset with only 4.85 M parameters and 38.19 G floating-point operations (FLOPs). The code will be released soon on the GitHub repository: https://github.com/dyl96/LTDNet