Sparse R-CNN OBB: Ship Target Detection in SAR Images Based on Oriented Sparse Proposals

Kamirul Kamirul,Odysseas Pappas,Alin Achim

2024-09-12

Abstract:We present Sparse R-CNN OBB, a novel framework for the detection of oriented objects in SAR images leveraging sparse learnable proposals. The Sparse R-CNN OBB has streamlined architecture and ease of training as it utilizes a sparse set of 300 proposals instead of training a proposals generator on hundreds of thousands of anchors. To the best of our knowledge, Sparse R-CNN OBB is the first to adopt the concept of sparse learnable proposals for the detection of oriented objects, as well as for the detection of ships in Synthetic Aperture Radar (SAR) images. The detection head of the baseline model, Sparse R-CNN, is re-designed to enable the model to capture object orientation. We also fine-tune the model on RSDD-SAR dataset and provide a performance comparison to state-of-the-art models. Experimental results shows that Sparse R-CNN OBB achieves outstanding performance, surpassing other models on both inshore and offshore scenarios. The code is available at: <a class="link-external link-http" href="http://www.github.com/ka-mirul/Sparse-R-CNN-OBB" rel="external noopener nofollow">this http URL</a>.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The main objective of this paper is to propose a new framework, Sparse R-CNN OBB, for rotational object detection in Synthetic Aperture Radar (SAR) images, specifically for ship detection. Specifically, this study addresses the following issues:

1. **Simplifying Model Design**: By utilizing sparse learnable proposals, the design and training process of the model is simplified. Traditional anchor-based methods require a large number of anchors to generate proposals, whereas Sparse R-CNN OBB uses only 300 sparse proposals.

2. **Improving Detection Performance**: By redesigning the detection head of the baseline model Sparse R-CNN to capture the orientation information of objects, the performance of detecting rotational objects (especially ships) in SAR images is improved.

3. **Experimental Validation**: Detailed experimental validation was conducted on the RSDD-SAR dataset, and comparisons were made with other state-of-the-art methods. The results show that Sparse R-CNN OBB performs excellently in various scenarios, particularly in ship detection in offshore and nearshore mixed environments.

In summary, this paper aims to develop an efficient and high-performance rotational object detection model by introducing the concept of sparse learnable proposals, which is particularly suitable for ship detection tasks in SAR images.