1.Zhejiang Provincial Key Laboratory of Estuary and Coast, Hangzhou 310020, China
2.College of Civil Engineering, Fuzhou University, Fuzhou ;350116, China
3.Key Laboratory of Sediment Sediment Science and Northern River Training, the Ministry of Water Resources, China Institute of;Water Resources and Hydropower Research, Beijing 100048, China
4.State Key Laboratory of Simulation and Regulation of Water Cycle in River;Basin, China Institute of Water Resources and Hydropower Research, Beijing 100048, China
5.Research Center for Yangtze River Ecological and;Environmental Engineering, China Three Gorges Corporation, Beijing 100038, China
Show less
文章历史+
Received
Accepted
Published
2023-09-24
Issue Date
2025-10-27
PDF (6554K)
摘要
推移质运动规律对河床演变具有重要影响,是河流动力学研究的重点和难点。本文开展中低水流条件下的推移质平衡输沙试验,将灰度相减方法和深度学习方法相结合,旨在提出一种优化的推移质运动颗粒识别算法,并在此基础上应用粒子跟踪测速技术(PTV)和卡尔曼滤波算法计算推移质颗粒运动轨迹,从而建立紊流相干结构与颗粒运动强度关系。为清晰捕捉运动距离较小颗粒,对YOLOv5(you only look once)目标检测模型网络结构进行卷积块改进、增加注意力机制和优化损失函数处理,以增强其在推移质颗粒识别任务中对极小目标的检测能力。结果表明:1)灰度相减方法可识别运动距离较大的颗粒,改进YOLOv5模型则能够更好识别运动距离较小的颗粒,通过合并优化两种方法,在本文设置的试验工况下可更准确识别推移质运动颗粒和运动轨迹;2)中低水流强度条件下,推移质泥沙颗粒运动强度受紊流相干结构影响沿水槽横向方向呈现间隔条带结构,并且随着水流强度增加,条带结构逐渐由密集变得稀疏,其宽度也由窄转宽,当水流强度继续增大时紊动掺混剧烈,条带结构遭到破坏逐渐消失;3)总体看来:泥沙颗粒运动多集中在水槽中间区域,边壁处运动较少,沙条带结构呈现中间宽两边窄的空间分布特征 ,表明条带结构是受紊流大尺度相干结构的影响而形成,而非二次流结构。
Abstract
Objective The movement of sediment particles significantly affects riverbed evolution, establishing it as a central concern and ongoing challenge in fluvial dynamics research. Although image processing methods provide efficient means for acquiring and analyzing data on sediment transport characteristics, their accuracy is often compromised by water waves, bubbles, and threshold errors. Therefore, continuous improvements and refinements remain essential to ensure the acquisition of accurate and reliable particle state data. This study integrates deep learning networks with existing image processing techniques to enable more precise and comprehensive identification of suspended sediment particles. It further investigates the relationship between turbulent coherent structures and the intensity of particle movement, clarifying the mechanism through which turbulent coherent structures influence sediment transport. Methods The optimization algorithm developed in this study aims to maximize the detection of moving particles, providing more accurate data to support understanding sediment transport patterns at the particle scale and their association with turbulent coherent structures. This research provides new insights for advanced measurement techniques and the exploration of sediment transport mechanisms. Bedload equilibrium sediment transport experiments are conducted under medium to low flow conditions (Θ = 0.052 to 0.071). High-speed cameras are utilized to capture images of bedload particles during water flow scouring processes. An optimized method for identifying bedload particle motion is proposed by combining the grayscale subtraction method with deep learning techniques. The grayscale subtraction method identifies regions of particle motion by calculating differences in grayscale values between consecutive frames and separately analyzing the centroids of moving particles in each frame. However, because this method depends solely on grayscale variations, it presents limitations in identifying regions with minor grayscale changes. The YOLOv5 (you only look once) method is designed to rapidly and accurately detect specific target objects and their locations in images after training on a sampled dataset. The YOLOv5 algorithm adopted in this study excels at detecting small targets and provides multi-scale detection, strong versatility, fast training, inference speeds, and adaptable fine-tuning capabilities. The YOLOv5 deep learning network structure is enhanced by improving convolutional blocks, incorporating attention mechanisms, and optimizing loss function processing, boosting the detector's overall performance in accurately capturing the motion of particles over short distances. The particle tracking velocimetry method and Kalman filtering algorithm are employed to calculate the trajectories of bedload particles. Results and Discussion The improved YOLOv5 model demonstrates significant enhancements in loss function handling, detection accuracy, and precision. The detection accuracy of the improved model for suspended sediment particles reaches 94.9%, with a 2.3% increase in average precision and respective gains of 1.1% and 1.0% in precision and recall rates. The weighted harmonic mean of the comprehensive verification index, F1 score, increases by two percentage points. This enhanced performance in practical detection surpasses that of the original YOLOv5 model. The number of observed particle chains increases following optimization by integrating the improved YOLOv5 model with the grayscale subtraction technique for detecting particle motion. Analyses of cumulative centroid counts and particle chain node counts reveal an ascending trend as the number of frames increases. The cumulative centroid count and particle chain node count obtained through the optimization method remain stable at approximately 59% and 80%, respectively, contrasting with the growth percentages of the individual methods. It is proposed that the formation of sediment particle motion bands is associated with Q2/Q4 bursting events of coherent turbulent structures based on the results of particle motion. During Q4 events, the average flow velocity exceeds that observed during Q2 events. Under identical water depth conditions, the shear force in the Q4 region surpasses that in the Q2 region, resulting in a higher concentration of sediment particles in the corresponding Q4 region. The bed surface structure exhibits convex grooves in the Q4 region and concave grooves in the Q2 region, extending across the entire bed surface in the spanwise direction of the channel. Characteristics of coherent turbulent structures provide a more comprehensive explanation for the mechanism underlying the formation of sediment transport belts. Conclusion This study concludes the following: 1) The grayscale subtraction technique effectively identifies particles with significant motion distances, while deep learning methods excel at recognizing particles with smaller motion distances. Through comparative analysis, data evaluation, and experimental observations, it becomes evident that the integrated algorithm, which combines both approaches, enhances the accuracy of bedload particle and trajectory identification under moderate to low flow conditions. 2)Under conditions of moderate to low flow intensity, the motion intensity of bedload sediment particles is influenced by coherent turbulent structures, resulting in a laterally banded structure. As flow intensity increases, the banded structure becomes sparser and wider. However, further intensification of the flow leads to vigorous turbulent mixing, which weakens the coherent turbulent structures and ultimately causes the banded structure to disappear. 3)Overall, sediment particle motion primarily concentrates in the central region of the channel, with reduced motion observed near the sidewalls due to lower flow velocities within the boundary layer. This observation aligns with practical scenarios and hydraulic theory. In addition, the morphology of the sediment-streaky structure typically exhibits a wider middle section and narrower sides, indicating that the formation of the banded structure is primarily influenced by large-scale coherent turbulent structures rather than secondary flow structures. This study introduces deep learning into conventional bedload particle motion recognition, improving the accuracy of bedload particle identification from a higher-resolution perspective. It addresses the challenge of detecting multiple moving sediment particles and provides a useful reference for research in fluvial dynamics.
本文在前人研究基础上,开展恒定均匀流推移质平衡输沙试验,借助高速摄像技术采集推移质运动图像。使用灰度相减方法获取运动距离较大颗粒,同时建立YOLOv5(you only look once)[27]深度学习模型,精确捕捉运动距离较小的运动颗粒。将两种结果进行合并后,使用PTV和卡尔曼滤波方法跟踪泥沙颗粒运动轨迹。在此基础上,研究推移质运动规律和分布特征,探讨紊流结构和泥沙颗粒运动强度带状结构耦合规律,为研究河床输沙提供参考依据。
式(1)~(2)中: Si 为位置i上构造的对称正定矩阵;为输入特征图第c个通道的局部区域Ωij 展开后的1维向量;wij 为i、j空间位置的关系权重; xi 、 xj 分别为输入特征图位置i、j的特征向量;Ωi 为以位置i为中心的滑动窗口邻域;C′为输入通道数;Kd,c 为卷积核权重参数 K 的元素, K,d为输出通道索引;Yh,w,d 为卷积输出在位置(h,w)、第d个通道的值。
①压缩(squeeze):通过局平均池化操作对每个特征图进行降维,将 X 的每个通道进行压缩,得到一个1×1×C的1维张量,即:
得到每个通道的压缩信息,生成对应的通道描述符。
②激发(excitation):将上一步得到的1维张量 z 首先输入一个全连接层,将其映射到一个激活函数前的低维空间:
式中: s 为中间变量, s ∈;Vec(·)表示将一个张量变为列向量;为修剪操作;为激活函数;W1为全连接层权重,W1∈。通过以上步骤,SE注意力模块可自适应地学习每个通道的重要性,对不同通道给出不同权重,提高模型对推移质颗粒特征的捕获。同时在主干网络(Backbone)的最后一层C3层前增加了SE模块,允许其以加权方式融合图像特征,提升网络性能。
在使用模型进行目标检测时,需要对模型进行评估,并选择适合自己需求的模型。以平均精度Mean average precision (mAP,记为)、召回率Recall(记为R)、精确率Precision(记为P)和P、R的加权调和平均数F1值(记为F1)为评价指标。mAP为AP的平均值,AP为所有召回率下的精度平均值(记为),即每个类别的精度召回率曲线(PR曲线)下方的面积,通常用于比较不同模型的性能。分别定义如下:
CameronS, NikoraV, WitzM.Entrainment of sediment particles by very large-scale motions[J].Journal of Fluid Mechanics,2020,888:A7. doi:10.1017/jfm.2020.24
[3]
GilbertG K, MurphyE C.The transportation of debris by running water[M].Fort Collins:US Geological Survey, 1914. doi:10.3133/pp86
[4]
EinsteinH A.The bed-load function for sediment transportation in open channel flows[R].Washington:USDA Technical Bulletin No.1026,1950.
[5]
BagnoldR A.An approach to the sediment transport problem from general physics[R].Washington:US Government Publishing Office,1966. doi:10.3133/pp422i
[6]
EngelundF.Closure to "hydraulic resistance of alluvial streams"[J].Journal of the Hydraulics Division,1967,93(4):287‒296. doi:10.1061/jyceaj.0001658
[7]
CrockerJ C, GrierD G.Methods of digital video microscopy for colloidal studies[J].Journal of Colloid and Interface Science,1996,179(1):298‒310. doi:10.1006/jcis.1996.0217
[8]
KeaneR D, AdrianR J.Theory of cross-correlation analysis of PIV images[J].Applied Scientific Research,1992,49(3):191‒215. doi:10.1007/bf00384623
[9]
AntonenkovD A, ShchodroA E.Using the PIV method to study the structure of the water flow and bottom sediment transport processes[J].Oceanology,2021,61(6):1064‒1071. doi:10.1134/s0001437021060187
[10]
HajisharifiA, MarchioliC, SoldatiA.Particle capture by dr-ops in turbulent flow[J].Physical Review Fluids,2021,6(2):02430. doi:10.1103/physrevfluids.6.024303
[11]
ElhimerM, PraudO, MarchalM,et al.Simultaneous PIV/PTV velocimetry technique in a turbulent particle-laden flow[J].Journal of Visualization,2017,20(2):289‒304. doi:10.1007/s12650-016-0397-z
[12]
WangLongwei, PanChong, LiuJianhua,et al.Ratio-cut ba-ckground removal method and its application in near-wall PTV measurement of a turbulent boundary layer[J].Measurement Science and Technology,2021,32(2):025302. doi:10.1088/1361-6501/abb483
[13]
DemiralD, AlbayrakI, TurowskiJ M,et al.Particle saltation trajectories in supercritical open channel flows:Roughness effect[J].Earth Surface Processes and Landforms,2022,47(15):3588‒3610. doi:10.1002/esp.5475
[14]
CampagnolJ, RadiceA, NokesR,et al.Lagrangian analysis of bed-load sediment motion:Database contribution[J].Jo-urnal of Hydraulic Research,2012,51(5):589‒596. doi:10.1080/00221686.2013.812152
[15]
LajeunesseE, MalvertiL, CharruF.Bed load transport in turbulent flow at the grain scale:Experiments and modeling[J].Journal of Geophysical Research:Earth Surface,2010, 115(F4):F04001. doi:10.1029/2009jf001628
[16]
HeymanJ.A study of the spatio-temporal behaviour of bed load transport rate fluctuations[D].Lausanne:École Po-lytechnique Fédérale de Lausanne,2014.
[17]
MiaoWei.Experimental research on bedload balance sediment transport based on image processing technology [D].Beijing:Tsinghua University,2018.
[18]
苗蔚.基于图像处理技术的推移质平衡输沙试验研究[D].北京:清华大学,2018.
[19]
MiaoWei, LiDanxun, ZhongQ.An image-based uninterrupted Lagrangian tracking method for measurement of sediment motion at grain scale[J].Applied Sciences,2020,10(13):460. doi:10.3390/app10134603
[20]
KlineS J, ReynoldsW C, SchraubF A,et al.The structure of turbulent boundary layers[J].Journal of Fluid Mechanics,1967,30(4):741‒77. doi:10.1017/s0022112067001740
[21]
NintoY, GarciaM H.Experiments on particle—Turbulence interactions in the near‒wall region of an open channel flow:Implications for sediment transport[J].Journal of Flu-id Mechanics,1996,326:285‒319. doi:10.1017/s0022112096008324
[22]
KarczI.Secondary currents and the configuration of a natural stream bed[J].Journal of Geophysical Research,1966,71(12):3109‒311. doi:10.1029/jz071i012p03109
[23]
NezuI, NakagawaH, JirkaG H.Turbulence in open-channel flows[J].Journal of Hydraulic Engineering,1994,120(10):1235‒1237. doi:10.1061/(asce)0733-9429(1994)120:10(1235)
[24]
WangZhiqian, ChengNiansheng.Secondary flows over artificial bed strips[J].Advances in Water Resources,2005,28(5):441‒450. doi:10.1016/j.advwatres.2004.12.008
[25]
YangShuqing, TanS K, WangXikun.Mechanism of secondary currents in open channel flows[J].Journal of Geophysical Research:Earth Surface,2012,117(F4):2012JF00-2510. doi:10.1029/2012jf002510
[26]
ZhongQiang, ChenQigang, WangHao,et al.Statistical ana-lysis of turbulent super-streamwise vortices based on observations of streaky structures near the free surface in the smooth open channel flow[J].Water Resources Research,2016,52(5):3563‒3578. doi:10.1002/2015wr017728
[27]
DuanYanchong, ChenQigang, LiDanxun,et al.Contributions of very large-scale motions to turbulence statistics in open channel flows[J].Journal of Fluid Mechanics,2020,892:A3. doi:10.1017/jfm.2020.174
[28]
RedmonJ, DivvalaS, GirshickR,et al.You only look once:Unified,real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:779‒788. doi:10.1109/cvpr.2016.91
[29]
RadiceA, MalavasiS, BallioF.Solid transport measurements through image processing[J].Experiments in Fluids,2006,41(5):721‒734. doi:10.1007/s00348-006-0195-9
[30]
ParkerG, KlingemanP C, McLeanD G.Bedload and size distribution in paved gravel-bed streams[J].Journal of the Hydraulics Division,1982,108(4):544‒571. doi:10.1061/jyceaj.0005854
[31]
WangHao.Study on coherent structure of turbulence and its influence on bed load motion characteristics[D].Beijing:Tsinghua University,2016.
[32]
王浩.紊流相干结构及对推移质运动特性影响的研究[D].北京:清华大学,2016.
[33]
WangHao, FanJieling, ZhongQiang,et al.Experimental study on sand strip structure of bedload bed based on high-speed camera technology [J].Journal of Hydraulic Engineering,2020,51(8):987‒996.
LeiFei, TangFeifei, LiShuhan.Underwater target detection algorithm based on improved YOLOv5[J].Journal of Marine Science and Engineering,2022,10(3):310. doi:10.3390/jmse10030310
[36]
SunkaraR, LuoTie.No more strided convolutions or pooling:A new CNN building block for low-resolution images and small objects[M]//Machine Learning and Knowledge Discovery in Databases.Cham:Springer Nature Switzerland,2023:443‒459. doi:10.1007/978-3-031-26409-2_27
[37]
HuJie, ShenLi, SunGang,et al.Squeeze-and-excitation networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(8):2011‒2023. doi:10.1109/tpami.2019.2913372
[38]
TongZanjia, ChenYuhang, XuZewei,et al.Wise‒IoU:Bounding box regression loss with dynamic focusing mechanism[EB/OL].(2023‒01‒24)[2023‒09‒01].
[39]
WillmarthW W, LuS S.Structure of the Reynolds stress near the wall[J].Journal of Fluid Mechanics,1972,55(1):65‒92. doi:10.1017/s002211207200165x
[40]
GualaM, TomkinsC D, ChristensenK T,et al.Vortex organization in a turbulent boundary layer overlying sparse roughness elements[J].Journal of Hydraulic Research,2012,50(5):465‒481. doi:10.1080/00221686.2012.729229