基于无人船的双分支解码轻量型分割网络研究
Research on dual branching decoding lightweight segmentation networks based on unmanned ships
刘丹 1张建杰1
作者信息
- 1. 新疆大学机械工程学院,乌鲁木齐 830000
- 折叠
摘要
为保证水面无人艇(USVs)进行水上任务时能够顺利航行,需要对河道信息进行精确的提取,因此,对河道语义分割的网络模型进行了研究.针对河道图像分割中类间不一致和类内不一致的问题,文中提出了分割网络DBDL-Net,网络中设计双分支解码结构和双重损失函数,分别把握语义信息和空间信息;同时在编码部分设计了多尺度残差的轻量模块,一方面减少参数,一方面以不同的比例捕捉特征信息.最后在USVIn-land数据集上对模型进行消融和对比实验,实验结果表明:DBDL-Net的精确度和平均交并比最终达到了93.619％和87.682％,与其他先进分割网络相比,DBDL-Net也具有更佳的综合表现.
Abstract
In order to ensure smooth navigation of unmanned surface vessels(USVs)for water missions,accurate extraction of river information is required,so a semantic segmentation network model of river is investigated.To ad-dress the problem of inter-class inconsistency and intra-class inconsistency in river image segmentation,a segmenta-tion network DBDL-Net is proposed in the paper,in which a double-branch decoding structure and a double loss function are designed to capture semantic and spatial information respectively;a lightweight module with multi-scale residuals is also designed in the coding part to reduce parameters on the one hand and capture feature information at different scales on the other.Finally,the model is ablated and compared with experiments on the USVInland dataset.The experimental results show that the accuracy and the mloU of DBDL-Net finally reach 93.619％and 87.682％,and DBDL-Net also has better overall performance compared with other advanced segmentation networks.
关键词
水面无人艇/DBDL-Net/双分支解码结构/双重损失函数/多尺度残差的轻量模块Key words
unmanned surface vessels/DBDL-Net/dual branch decoding structure/double loss function/light-weight module with multiscale residuals引用本文复制引用
基金项目
新疆维吾尔自治区重点研发计划(2022B02038)
出版年
2024
NETL
NSTL
万方数据