[1]王子垚,王梓源,钱保如,等.基于轻量化YOLOv8s模型的无人机河道漂浮物实时检测方法研究[J].江苏水利,2025,(01):21-25.
　WANG Ziyao,WANG Ziyuan,QIAN Baoru,et al.Research on real-time detection method of floating debris in river using UAV based on lightweight YOLOv8s model[J].JIANGSU WATER RESOURCES,2025,(01):21-25.
点击复制

基于轻量化YOLOv8s模型的无人机河道漂浮物实时检测方法研究()
分享到：

《江苏水利》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年01期
页码:
21-25
栏目:
水利信息化
出版日期:
2025-01-15

文章信息/Info

Title:
Research on real-time detection method of floating debris in river using UAV based on lightweight YOLOv8s model
文章编号:
1007-7839（2025）01-0021-0005
作者:
王子垚;  王梓源;  钱保如;  陆冠臣;  王海蓉
（江苏省骆运水利工程管理处，江苏 宿迁 223800）
Author(s):
WANG Ziyao;  WANG Ziyuan;  QIAN Baoru;  LU Guanchen;  WANG Hairong
（Luoyun Hydraulic Project Management Division of Jiangsu Province, Suqian 223800, China）
关键词:
漂浮物识别; 无人机; 轻量化; 实时检测; YOLOv8s
Keywords:
floating debris identification;  UAV;  lightweight;  real-time detection;  YOLOv8s
分类号:
TP391.4
文献标志码:
B
摘要:
为解决河道漂浮物检测中微小漂浮物的识别问题，提出结合YOLOv8s、LSKA和AKConv的改进模型，将YOLOv8s的参数从11.1M降至8.3M，并在“中国水科院水面漂浮物数据集”上将mAP50从68.0%提升至77.9%。结果表明，该方法能提高检测精度，增强遮挡识别能力，使网络更适应河道复杂环境，有效提升无人机在河道中的实时检测能力，为水利管理提供了高效解决方案。
Abstract:
To solve the problem of identifying small floating objects in river floating object detection, an improved model combining YOLOv8s, LSKA, and AKConv is proposed. The improved model reduces the parameters of YOLOv8s from 11.1M to 8.3M, and increases mAP50 from 68.0% to 77.9% on the "Chinese Academy of Water Surface Floating Object Dataset". The results show that this method can improve detection accuracy, enhance occlusion recognition ability, make the network more adaptable to complex river environments, effectively enhance the real-time detection capability of UAV in rivers, and provide an efficient solution for water management.

参考文献/References:

[1]汤明文，戴礼豪，林朝辉，等. 无人机在电力线路巡视中的应用[J]. 中国电力，2013（3）：35-38.
[2]CHENG Y，XU H，LIU Y. Robust small object detection on the water surface through fusion of camera and millimeterwave radar[J]. Proceedings of the IEEE/CVF internationalconference on computer vision，2021，11（4）：15263-15272.
[3]王乾胜，展勇忠，邹宇. 基于改进Yolov5n的无人机对地面军事目标识别算法[J]. 计算机测量与控制，2024，32（6）：189-197，226.
[4]张华卫，张文飞，蒋占军，等. 引入上下文信息和AttentionGate的GUS-YOLO遥感目标检测算法[J]. 计算机科学与探索，2024，18（2）:453-464.
[5]杨明祥，乔广超，王浩，等. 水面漂浮物数据集（IWHR_AI_Lable_Floater_V1）[DB/OL]. 中国水利水电 科学研究院，2023. http://123.56.14.89:8008/wfdownload/.

相似文献/References:

[1]张宁宁,丁旭东,赵 鹏.无人机航空摄影测量在河道地形图测绘中的应用[J].江苏水利,2018,(12):31.
　ZHANG Ningning,DING Xudong,ZHAO Peng.Application of UAV aerial photogrammetry in river topographic mapping[J].JIANGSU WATER RESOURCES,2018,(01):31.
[2]赵建华,叶文,阮哲伟,等.无人机视觉测流技术在河流流量监测中的应用[J].江苏水利,2022,(07):37.
　ZHAO Jianhua,YE Wen,RUAN Zhewei,et al.Application of UAV visual flow velocimetry in river flow monitoring[J].JIANGSU WATER RESOURCES,2022,(01):37.
[3]曹文星,张炜杰,游天宇,等.无人机搭载双光相机探测堤防渗漏试验分析[J].江苏水利,2022,(10):63.
　CAO Wenxing,ZHANG Weijie,YOU Tianyu,et al.Test analysis on dike-leakage-detection by using UAV equipped with visible-and-infrared-light-lens[J].JIANGSU WATER RESOURCES,2022,(01):63.
[4]朱大栋,朱昊*,王远,等.蓄滞洪区洪灾避险迁移路径规划研究[J].江苏水利,2023,(02):65.
　ZHU Dadong,ZHU Hao,WANG Yuan,et al.Study on flood avoidance migration path planning in flood storage and detention area[J].JIANGSU WATER RESOURCES,2023,(01):65.
[5]朱昊,张引,王茂枚,等.基于无人机技术与数学模型的蓄滞洪区洪水演进分析[J].江苏水利,2023,(06):49.
　ZHU Hao,ZHANG Yin,WANG Maomei,et al.Analysis on flood routing in flood storage and detention area based on UAV and mathematical model[J].JIANGSU WATER RESOURCES,2023,(01):49.
[6]方卫华,杨浩东,谢双双,等.面向堤坝监测的无人机自动巡检系统设计[J].江苏水利,2024,(02):1.
　FANG Weihua,YANG Haodong,XIE Shuangshuang,et al.Design of UAV automatic inspection system for dam monitoring[J].JIANGSU WATER RESOURCES,2024,(01):1.
[7]赵思曾,康飞,何丽娜,等.小型水库大坝无人机巡检和行人定位方法[J].江苏水利,2024,(06):12.
　ZHAO Sizeng,KANG Fei,HE Lina,et al.Unmanned aerial vehicle inspection and pedestrian location method for small reservoir dams[J].JIANGSU WATER RESOURCES,2024,(01):12.

备注/Memo

备注/Memo:
收稿日期：2024-09-23
作者简介：王子垚（1996—），男，硕士，主要从事水利工程生产运行工作。E-mail：1214436643@qq.com
更新日期/Last Update: 2025-01-15