Design of Forward-Looking Sonar System for Real-Time Image Segmentation with Light Multi-Scale Attention Net

计算机科学 分割 声纳 人工智能 特征提取 特征(语言学) 计算机视觉 图像分割 交叉口(航空) 判别式 深度学习 模式识别(心理学) 工程类 语言学 哲学 航空航天工程
作者
Dongdong Zhao,Hongchao Zhou,Peng Chen,Yingtian Hu,Weibo Mao,Yuanjie Dang,Ronghua Liang
出处
期刊:IEEE Transactions on Instrumentation and Measurement [Institute of Electrical and Electronics Engineers]
卷期号:: 1-1
标识
DOI:10.1109/tim.2023.3341127
摘要

Forward-looking sonar is a commonly used underwater detection device. However, due to the complex underwater environment, small target areas, and blurred features, the detection accuracy is poor and not conducive to quickly finding the desired targets. In recent years, deep learning has developed rapidly, and semantic segmentation network models have shown strong segmentation performance, with enormous application potential in forward-looking sonar image segmentation. A lightweight multi-scale attention network is proposed in this paper for real-time semantic segmentation of forward-looking sonar system. The proposed model ensures the training accuracy and inference speed of the semantic segmentation network through the structural reparameterization module and improves the feature extraction ability of the network by fusing multi-scale information through the multi-scale feature attention gate. Compared with other models, the proposed model achieved good results on semantic segmentation datasets and data collected from self-developed devices. The mean intersection over union (mIoU) reached 0.734 on a public dataset, outperforming existing mainstream models. Moreover, the inference speed can reach 1250 FPS on 2080Ti. An embedded forward-looking sonar host computer is designed based on Atlas200, and the proposed network can run up to 87 FPS on the embedded host computer, with a frame rate of 9.5 FPS per unit of power consumption, which is better than that of mobile phones and X86 computers. The forward-looking sonar system constructed based on this achieved satisfactory results in engineering practice.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
大幅提高文件上传限制,最高150M (2024-4-1)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
科研通AI2S应助zxvcbnm采纳,获得10
刚刚
Always完成签到,获得积分10
1秒前
1秒前
zzz发布了新的文献求助30
2秒前
2秒前
混子发布了新的文献求助10
2秒前
盼盼完成签到,获得积分10
4秒前
5秒前
6秒前
7秒前
穆紫应助kk采纳,获得10
7秒前
搜集达人应助zzz采纳,获得30
7秒前
研友_VZG7GZ应助阿三采纳,获得10
8秒前
盼盼发布了新的文献求助10
9秒前
虚拟的姒发布了新的文献求助20
9秒前
爆米花应助吃人陈采纳,获得10
10秒前
zxvcbnm发布了新的文献求助10
10秒前
缥缈伟祺完成签到,获得积分20
11秒前
11秒前
杀手爱吃小熊饼干完成签到,获得积分20
11秒前
DianaRang发布了新的文献求助10
12秒前
12秒前
12秒前
初级小白发布了新的文献求助10
13秒前
14秒前
14秒前
14秒前
15秒前
月光取暖发布了新的文献求助10
16秒前
狂野元枫发布了新的文献求助10
16秒前
huo发布了新的文献求助10
16秒前
ningning完成签到 ,获得积分10
16秒前
zzz完成签到,获得积分20
16秒前
鹿不羁完成签到 ,获得积分10
16秒前
whz完成签到,获得积分10
17秒前
17秒前
李健应助橙子采纳,获得10
18秒前
1233发布了新的文献求助10
18秒前
mitty完成签到,获得积分10
20秒前
20秒前
高分求助中
Sustainability in Tides Chemistry 2000
Bayesian Models of Cognition:Reverse Engineering the Mind 800
Essentials of thematic analysis 700
A Dissection Guide & Atlas to the Rabbit 600
Very-high-order BVD Schemes Using β-variable THINC Method 568
Mantiden: Faszinierende Lauerjäger Faszinierende Lauerjäger 500
PraxisRatgeber: Mantiden: Faszinierende Lauerjäger 500
热门求助领域 (近24小时)
化学 医学 生物 材料科学 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 基因 遗传学 催化作用 物理化学 免疫学 量子力学 细胞生物学
热门帖子
关注 科研通微信公众号,转发送积分 3124803
求助须知:如何正确求助?哪些是违规求助? 2775148
关于积分的说明 7725553
捐赠科研通 2430633
什么是DOI,文献DOI怎么找? 1291291
科研通“疑难数据库(出版商)”最低求助积分说明 622121
版权声明 600328