[1]张思思,滑文强. 基于时频双域协同与语义增强的复杂水域漂浮物检测方法[J].机械与电子,2026,44(03):47-54.
　ZHANG Sisi,HUA Wenqiang. A Dual-domain Synergistic and Semantically Enhanced Method for Floating Object Detection in Complex Water Environments[J].Machinery & Electronics,2026,44(03):47-54.
点击复制

 基于时频双域协同与语义增强的复杂水域漂浮物检测方法()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
44
期数:
2026年03期
页码:
47-54
栏目:
智能检测
出版日期:
2026-03-25

文章信息/Info

Title:
 A Dual-domain Synergistic and Semantically Enhanced Method for Floating Object Detection in Complex Water Environments
文章编号:
1001-2257(2026)03-0047-08
作者:
 张思思1; 滑文强2
 (1.西安航空职业技术学院自动化工程学院,陕西 西安 710089;
2.西安邮电大学计算机学院,陕西 西安 710061)
Author(s):
 ZHANG Sisi1; HUA Wenqiang2
 (1.School of Automation Engineering,Xi’an Aeronautical Polytechnic Institute,Xi’an 710089,China;
2.School of Computer Science and Technology,Xi’an University of Posts and Telecommunications,Xi’an 710061,China)
关键词:
 ; 目标检测; 空频自适应; 小波变换; Inner IoU; 水面漂浮物
Keywords:
 ; object detection; spatial frequency adaptive; wavelet transform; Inner IoU; water surface floating objects
分类号:
TP391.4
文献标志码:
A
摘要:
 为解决复杂水域环境下漂浮物检测中存在的尺度差异显著、背景干扰强烈及小目标漏检等问题,提出一种空间频域增强网络(SFE Net)。首先,提出空频自适应卷积,通过空间频率分流与自适应融合策略,突破了传统卷积固定感受野的局限,实现了空间上下文与频率纹理特征的协同提取,显著增强了对多尺度目标的感知能力。其次,设计小波频率解耦模块,利用Haar小波变换将特征解耦为高频细节与低频轮廓,并通过双路注意力机制分别强化边缘纹理与全局语义表征,有效抑制了水面波纹与反光等复杂背景噪声。最后,引入Inner IoU 边界框回归损失函数,通过引入辅助边框与比例因子控制,优化样本在特征空间的内部相似性度量,解决了小目标定位中IoU 敏感度不足的问题,提高了边界框回归的收敛速度与定位精度。在FloW+数据集上的实验结果表明,该方法的mAP@50达到91.4%,相比原始YOLOv8n提升8.7百分点;同时模型参数量减少36.7%,推理速度达到214 帧/s,在保证实时性的同时显著提升了检测性能。
Abstract:
To address the issues of significant scale variations,strong background interference,and high miss detection rates of small objects in floating debris detection under complex water environments,this paper proposes a Spatial Frequency Enhancement Network (SFE Net).First,a Spatial Frequency Adaptive Convolution is proposed.By employing a spatial frequency split strategy and an adaptive fusion scheme,it overcomes the limitation of the fixed receptive fields in traditional convolution,achieving the collaborative extraction of spatial contextual and frequency texture features,thereby significantly enhancing the perception capability for multi scale objects.Second,a Wavelet Frequency Decoupling Module is designed,which decouples features into high frequency details and low frequency contours using Haar wavelet transform,and employs dual path attention mechanisms to separately enhance edge textures and global semantic representations,effectively suppressing complex background noise such as water ripples and reflections.Finally,an Inner IoU bounding box regression loss function is introduced.By incorporating auxiliary bounding boxes and controlling scaling factors,it optimizes the internal similarity measurement of samples in feature space,addressing the problem of insufficient IoU sensitivity in small object localization,and improving the convergence speed and localization accuracy of bounding box regression.Experimental results on the FloW+ dataset demonstrate that the proposed method achieves a mean Average Precision of 91.4%,representing an improvement of 8.7 percentage points over the original YOLOv8n.Meanwhile,the model parameters are reduced by 36.7%,and the inference speed reaches 214 frames per second,significantly enhancing detection performance while ensuring real time capability.

参考文献/References:

 [1] TONG K,WU Y Q,ZHOU F.Recent advances in small object detection based on deep learning:a review[J].Image and vision computing,2020,97:103910.
[2] 徐守坤,龚研,李宁.FA YOLOv7:水面垃圾小目标检测算法[J].计算机工程与设计,2025,46(8):2328 2334.
[3] JIN X L,NIU P W,LIU L F.A GMM based segmentation method for the detection of water surface floats [J].IEEE Access,2019,7:119018 119027.
[4] STRICELJ A,KACIC Z.Detection of objects on waters’ surfaces using CEIEMV method[J].Computers and electrical engineering,2015,46:511 527.
[5] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real time object detection[C]∥2016 IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE,2016:779 788.
[6] ZHANG L L,WEI Y X,WANG H B,et al.Real time detection of river surface floating object based on improved RefineDet[J].IEEE Access,2021,9:81147 81160.
[7] 陈任飞,彭勇,吴剑,等.基于改进SSD 模型和自适应滤波算法的水面漂浮物目标检测跟踪方法[J].工程科学与技术,2023,55(4):119 129.
[8] 朱宇凡,杨吉凯,李威.改进YOLOv12的水面垃圾检测方法[J].计算机工程与应用,2026,62(1):101 111.
[9] ZHANG L L,XIE Z Q,XU M Q,et al.EYOLOv3:an efficient real time detection model for floating object on river[J].Applied sciences,2023,13(4):2303.
[10] 何君尧,王文胜,韩宜航.结合YOLOv8与双目测距算法的水面漂浮垃圾检测定位系统设计[J].现代电子技术,2024,47(20):1 7.
[11] LI N,WANG M L,YANG G C,et al.DENS YOLOv6:a small object detection model for garbage detection on water surface[J].Multimedia tools and applications,2023,83:55751 55771.
[12] CHENG Y W,ZHU J N,JIANG M X,et al.FloW:a dataset and benchmark for floating waste detection in inland waters[C]∥2021 IEEE/CVF International Conference on Computer Vision.New York:IEEE,2021:10953 10962.
[13] LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot MultiBox detector[C]∥Computer Vision ECCV 2016,2016:21 37.
[14] FENG Y F,HUANG J G,DU S Y,et al.Hyper YOLO:when visual object detection meets hypergraph computation[J].IEEE Transactions on pattern analysis and machine intelligence,2024,47 (4):2388 2401.

相似文献/References:

[1]雷永涛,周 权,张静静. 基于印制电路连接器变形问题的分析和设计改进[J].机械与电子,2019,(01):7.
　,analysis and design improvement of deformation problem based on printed circuit connector[J].Machinery & Electronics,2019,(03):7.
[2]孟繁萃１,张 达２,张 明１. 自然风力对绝缘子结构的影响[J].机械与电子,2019,(01):3.
　,the influence of natural wind on insulator structure[J].Machinery & Electronics,2019,(03):3.
[3]任智勇１,邓晓政１,马丁峰２. 一种发动机试飞数据快速处理方法[J].机械与电子,2019,(01):11.
　,A Fast Processing Method of Multi-subject Engine Flight Test Data[J].Machinery & Electronics,2019,(03):11.
[4]冷雪锋,吴正明. 基于ＳＴＣ１２Ｃ５Ａ６０Ｓ２的１００Ｗ 小型风力发电系统设计[J].机械与电子,2019,(01):15.
　.The Design of 100W Small Wind Power Generation System Based on STC12C5A60S2[J].Machinery & Electronics,2019,(03):15.
[5]查金水,朱志远,邓友银,等. 某大型相控阵雷达测试系统结构设计[J].机械与电子,2019,(01):19.
　,,et al.Structural Design of a Large Phased Array Radar Test System[J].Machinery & Electronics,2019,(03):19.
[6]万 欢,李伟达,李 娟. 一种面向上肢康复的便携式电刺激器设计[J].机械与电子,2019,(01):33.
　,Design of Portable Electric Stimulator for Upper Limb Rehabilitation[J].Machinery & Electronics,2019,(03):33.
[7]郑锦汤１,李玉忠１,２. 紧凑型纯电动汽车动力系统匹配与性能仿真[J].机械与电子,2019,(01):38.
　,Matching and Performance Simulation of Dynamic Systems of Compact Electric Vehicle[J].Machinery & Electronics,2019,(03):38.
[8]吕明珠１,２,苏晓明 １,等. 改进粒子群算法优化的支持向量机在滚动轴承故障诊断中的应用[J].机械与电子,2019,(01):42.
　,,et al.Application of SVM Optimized by IPSO in Rolling Bearing Fault Diagnosis[J].Machinery & Electronics,2019,(03):42.
[9]韩雪莹１,张佳炜２,刘 梦１,等. 集中控制协调多个并网逆变器的谐波补偿[J].机械与电子,2019,(01):49.
　,,et al.Centralized Control Coordination of Harmonic Compensation for Multiple Grid-connected Inverters[J].Machinery & Electronics,2019,(03):49.
[10]蒋 波１,３,周自强２,等. 基于ＰＬＣ的报废汽车拆解线作业控制系统[J].机械与电子,2019,(01):54.
　,,et al.PLC-based ELV Disassembly Line Operation Control System[J].Machinery & Electronics,2019,(03):54.

备注/Memo

备注/Memo:
 收稿日期:2025-12-12
基金项目:西安航空职业技术学院2023年度校级科研计划项目(23XHZK-15)
作者简介:张思思 (1989-),女,陕西西安人,讲师,研究方向为人工智能、嵌入式系统;滑文强 (1987-),男,陕西西安人,讲师,研究方向为机器学习。
更新日期/Last Update: 2026-04-29