[1]高睿,范子武,刘国庆,等.面向河湖监管的无人机视频全景融合技术研究与应用[J].江苏水利,2026,(06):46-51.
 GAO Rui,FAN Ziwu,LIU Guoqing,et al.Research and application of UAV video panoramic fusion technology for river and lake supervision[J].JIANGSU WATER RESOURCES,2026,(06):46-51.
点击复制

面向河湖监管的无人机视频全景融合技术研究与应用()

《江苏水利》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2026年06期
页码:
46-51
栏目:
水利信息化
出版日期:
2026-07-02

文章信息/Info

Title:
Research and application of UAV video panoramic fusion technology for river and lake supervision
文章编号:
1007-7839(2026)06-0046-0006
作者:
高睿12范子武12刘国庆12洪磊12王梓豪12
(1.南京水利科学研究院,江苏 南京 210029;2.水利部太湖流域水治理重点实验室,江苏 南京 210029)
Author(s):
GAO Rui12 FAN Ziwu12 LIU Guoqing12 HONG Lei12 WANG Zihao12
(1.Nanjing Hydraulic Research Institute, Nanjing 210029, China;2.Key Laboratory of the Taihu Basin Water Treatment of Ministry of Water Resouces, Nanjing 210029, China)
关键词:
多源视频融合无人机监测畸变矫正视频拼接河湖监管
Keywords:
multi-source video fusion UAV monitoring distortion correction video stitching river and lake supervision
分类号:
TV213.4
文献标志码:
A
摘要:
针对高空河湖监管中视角受限、广角畸变及复杂水面环境下配准不稳等问题,研究并实现了一种面向河湖监管的高空自由视角视频全景融合技术。首先针对广角镜头造成的几何失真,利用河岸线等天然几何特征进行自标定校正,为影像对齐奠定几何基础。其次,针对水面纹理稀疏导致的配准失效,采用ORB特征提取结合RANSAC算法剔除外点,通过求解最优单应性矩阵实现像素级精确配准。最后,为解决无人机姿态抖动引起的视频跳变及色差,引入动态配准机制并配合分层融合策略,实现了时空连贯、无缝的动态全景监测。在锦溪镇典型水乡的实验验证表明,该技术能显著提升巡河效率,为蓝藻识别与智慧管护提供了有效的数字化支撑。
Abstract:
To address the challenges in high-altitude river and lake supervision, such as limited field of view, severe wide-angle lens distortion, and unstable registration in complex water environments, a free-viewpoint panoramic video fusion technology for UAV-based monitoring is researched and implemented. First, to correct the geometric distortion caused by wide-angle UAV lenses, a self-calibration method utilizing natural geometric constraints—such as riverbank lines—is employed to establish an accurate foundation for image alignment. Second, to overcome registration failure in texture-sparse water surfaces, the ORB (Oriented FAST and Rotated BRIEF) algorithm combined with RANSAC (Random Sample Consensus) is used to eliminate outliers and solve for the optimal homography matrix, achieving pixel-level precision. Finally, to eliminate video jitter and color differences caused by UAV attitude fluctuations and photometric differences, a dynamic registration mechanism and layered blending strategy are introduced to generate seamless, spatiotemporally coherent panoramic monitoring results. Experimental validation in the typical river-lake network of Jinxi Town demonstrates that the proposed method significantly enhances river inspection efficiency and provides effective digital support for algal bloom detection and intelligent water management.

参考文献/References:

[1]蔡阳. 以数字孪生流域建设为核心构建具有“四预”功能智慧水利体系[J]. 中国水利,2022(20):2-6,60.
[2]李志安. 广角图像畸变矫正及双光谱视频中游泳池人员实时检测和跟踪[D]. 济南:济南大学,2024.
[3]SIMARRO G,CALVETE D,PLOMARITIS T A,et al. The influence of camera calibration on nearshore bathymetryestimation from UAV videos[J]. Remote Sensing,2021,13(1):150.
[4]LIAO K,NIE L,HUANG S,et al. Deep learning for cameracalibration and beyond:A survey[J]. arXiv preprint arXiv,2023(2303):10559.
[5]RUBLEE E,RABAUD V,KONOLIGE K,et al. ORB:Anefficient alternative to SIFT or SURF[C]//2011 Interna- tional conference on computer vision. Ieee,2011:2564-2571.
[6]DING Y,VAVRA V,BHAYANI S,et al. Fundamental matrix estimation using relative depths[C]//European Conference on Computer Vision. Cham:Springer NatureSwitzerland,2024:142-159.
[7]ZHANG H,SHEN H,YUAN Q,et al. Multispectral andSAR image fusion based on Laplacian pyramid and sparserepresentation[J]. Remote Sensing,2022,14(4):870.
[8]LI Q,MA Z,LI J,et al. Overview of the research status ofintelligent water conservancy technology system[J]. Applied Sciences,2024,14(17):7809.
[9]LI W,MA Z,LI J,et al. Digital twin smart water conservancy:Status,challenges,and prospects[J]. Water,2024,16(14):2038.
[10]闫慕云. 基于图像特征匹配的多路摄像头视频拼接算法研究[D]. 西安:西安石油大学,2024.
[11]崔鹏. 多目视频实时拼接技术研究[D]. 天津:中国民航大学,2023.
[12]KOUTALAKIS P,TZORAKI O,ZAIMES G. UAVs for hydrologic scopes:Application of a low-cost UAV to estimatesurface water velocity by using three different image-basedmethods[J]. Drones,2019,3(1):14.
[13]李昌令. 面向三维场景的多路视频拼接投影融合方法研究[D]. 西安:西安电子科技大学,2024.
[14]程丹丹,钮月磊,刘创. 基于数字孪生的智慧水利系统的设计与应用[J]. 现代信息科技,2025,9(2):105-109.
[15]ACHARYA B S,BHANDARI M,BANDINI F,et al. Unmanned aerial vehicles in hydrology and water manage-ment:Applications,challenges,and perspectives[J]. WaterResources Research,2021,57(11):e2021WR029925.
[16]HAMZAH A,AQLAN F,BAIDYA S. Drone-based digitaltwins for water quality monitoring:A systematic review[J]. Digital Twins and Applications,2024,1(2):131-160

备注/Memo

备注/Memo:
收稿日期:2026-03-14
基金项目:重点研发计划(2024YFC3210800);广西科技重大专项(桂科AA23062053);南京水利科学研究院研究生学位论文基金(Yy125005)
作者简介:高睿(2001—),男,硕士研究生,主要从事水利水电工程研究。E-mail:2017170689@qq.com
通信作者:范子武(1970—),男,正高级工程师,博士,主要从事洪灾风险分析、城市水力学、河湖治理、水利信息化等研究工作。E-mail:zwfan@nhri.cn
更新日期/Last Update: 2026-06-01