<p>Water surface floating waste collection using uncrewed surface vehicles (USVs) is essential for combating pollution caused by floating waste. However, the unique challenges of water surface images collected by vision sensors, such as small object sizes, wave disturbances, light reflections, and shoreline shadows, significantly reduce the effectiveness of existing object detection methods for identifying floating objects. In light of the above issues, we propose a novel network named SEBR-YOLOv8n. This method adopts an RT-DETR decoder based on Transformer architecture, combined with our proposed bidirectional weighted feature pyramid SE-BiFPN with small target enhancement, to significantly improve the detection performance of small objects from the USV perspective by enhancing the network’s information fusion capability. In addition, the Inner-SIoU was introduced to the network to accelerate bounding box regression and enhance the object detection capabilities. Experimental results show that SEBR-YOLOv8n (the model size is 5.0 M) can achieve the <InlineEquation ID="IEq1"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50}\)</EquationSource></InlineEquation> of 91.2% and the <InlineEquation ID="IEq2"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50-95}\)</EquationSource></InlineEquation> of 49.9% on the FloW-Img dataset, representing improvements of 6.2% in <InlineEquation ID="IEq3"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50}\)</EquationSource></InlineEquation> and 4.2% in <InlineEquation ID="IEq4"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50-95}\)</EquationSource></InlineEquation> compared to YOLOv8n. Furthermore, on Floating Waste-I dataset, the <InlineEquation ID="IEq5"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50}\)</EquationSource></InlineEquation> can reach 92.4% and the <InlineEquation ID="IEq6"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50-95}\)</EquationSource></InlineEquation> can reach 50.1%, validating the generalization of SEBR-YOLOv8n. We also use Image-enhancement projects to synthesize samples and further validate the performance of the algorithm. The <InlineEquation ID="IEq7"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50}\)</EquationSource></InlineEquation> and <InlineEquation ID="IEq8"><EquationSource Format="TEX">\(\:\text{m}\text{A}{\text{P}}_{50-95}\)</EquationSource></InlineEquation> of SEBR-YOLOv8n on the dataset containing synthetic samples reached 95.5% and 58.2%, respectively. The results show that SEBR-YOLOv8n can achieve high-precision detection of floating objects on the water surface while maintaining a lightweight structure.</p>

Lightweight detection network based on bidirectional weighted feature fusion with small target enhancement for USVs

期刊论文
出版模式：

开放获取
发表日期：

2026年6月21日

查看全文

Yong Li,
Dehang Lian,
Jialong Du,
Chunning Bu,
Dongxu Gao

摘要

Water surface floating waste collection using uncrewed surface vehicles (USVs) is essential for combating pollution caused by floating waste. However, the unique challenges of water surface images collected by vision sensors, such as small object sizes, wave disturbances, light reflections, and shoreline shadows, significantly reduce the effectiveness of existing object detection methods for identifying floating objects. In light of the above issues, we propose a novel network named SEBR-YOLOv8n. This method adopts an RT-DETR decoder based on Transformer architecture, combined with our proposed bidirectional weighted feature pyramid SE-BiFPN with small target enhancement, to significantly improve the detection performance of small objects from the USV perspective by enhancing the network’s information fusion capability. In addition, the Inner-SIoU was introduced to the network to accelerate bounding box regression and enhance the object detection capabilities. Experimental results show that SEBR-YOLOv8n (the model size is 5.0 M) can achieve the \(\:\text{m}\text{A}{\text{P}}_{50}\) of 91.2% and the \(\:\text{m}\text{A}{\text{P}}_{50-95}\) of 49.9% on the FloW-Img dataset, representing improvements of 6.2% in \(\:\text{m}\text{A}{\text{P}}_{50}\) and 4.2% in \(\:\text{m}\text{A}{\text{P}}_{50-95}\) compared to YOLOv8n. Furthermore, on Floating Waste-I dataset, the \(\:\text{m}\text{A}{\text{P}}_{50}\) can reach 92.4% and the \(\:\text{m}\text{A}{\text{P}}_{50-95}\) can reach 50.1%, validating the generalization of SEBR-YOLOv8n. We also use Image-enhancement projects to synthesize samples and further validate the performance of the algorithm. The \(\:\text{m}\text{A}{\text{P}}_{50}\) and \(\:\text{m}\text{A}{\text{P}}_{50-95}\) of SEBR-YOLOv8n on the dataset containing synthetic samples reached 95.5% and 58.2%, respectively. The results show that SEBR-YOLOv8n can achieve high-precision detection of floating objects on the water surface while maintaining a lightweight structure.