DiffDet4SAR: Diffusion-based Aircraft Target Detection Network for SAR Images

Zhou Jie,Xiao Chao,Peng Bo,Liu Zhen,Liu Li,Liu Yongxiang,Li Xiang

DOI: https://doi.org/10.1109/LGRS.2024.3386020

2024-04-17

Abstract:Aircraft target detection in SAR images is a challenging task due to the discrete scattering points and severe background clutter interference. Currently, methods with convolution-based or transformer-based paradigms cannot adequately address these issues. In this letter, we explore diffusion models for SAR image aircraft target detection for the first time and propose a novel \underline{Diff}usion-based aircraft target \underline{Det}ection network \underline{for} \underline{SAR} images (DiffDet4SAR). Specifically, the proposed DiffDet4SAR yields two main advantages for SAR aircraft target detection: 1) DiffDet4SAR maps the SAR aircraft target detection task to a denoising diffusion process of bounding boxes without heuristic anchor size selection, effectively enabling large variations in aircraft sizes to be accommodated; and 2) the dedicatedly designed Scattering Feature Enhancement (SFE) module further reduces the clutter intensity and enhances the target saliency during inference. Extensive experimental results on the SAR-AIRcraft-1.0 dataset show that the proposed DiffDet4SAR achieves 88.4\% mAP$_{50}$, outperforming the state-of-the-art methods by 6\%. Code is availabel at \href{

Image and Video Processing,Signal Processing

What problem does this paper attempt to address?

The paper attempts to address the problem of detecting aircraft targets in Synthetic Aperture Radar (SAR) images. Specifically, the paper points out that current convolution-based or transformer-based methods cannot adequately address the following challenges:

1. **Diversity in Aircraft Target Sizes**: Different types of aircraft targets and different resolutions of images in SAR lead to variations in target sizes, making it easy to miss small-sized or weakly scattered aircraft targets when using fixed receptive field techniques for feature extraction.
2. **Discrete Scattering Points**: The scattering characteristics of aircraft targets are discrete and discontinuous, with incomplete structures and weak correlations between components, making it difficult for detectors to achieve accurate predictions.
3. **Complex Scenes and Strong Background Interference**: In real-world scenarios, buildings, vehicles, and other metal structures exhibit strong scattering characteristics similar to aircraft targets, making target detection challenging.

To address these challenges, the paper explores the application of diffusion models in SAR image aircraft target detection for the first time and proposes a novel diffusion-based aircraft target detection network (DiffDet4SAR). This method effectively addresses the aforementioned issues by mapping the SAR aircraft target detection task to a denoising diffusion process from noisy bounding boxes to real bounding boxes. Additionally, the paper designs a Scattering Feature Enhancement (SFE) module to reduce background clutter intensity and enhance target saliency. Experimental results show that DiffDet4SAR achieves 88.4% mAP 50 on the SAR-AIRcraft-1.0 dataset, which is a 6% improvement over the existing best methods.