针对无人机公路日常巡检任务中检测目标小、背景复杂、计算量大导致的检测精度不足、实时性差等问题,该文提出一种基于YOLOv11n改进的无人机公路日常巡检算法YOLOv11-UR。该算法将MLCA注意力机制加入主干网络,融合通道与空间维度信息,...针对无人机公路日常巡检任务中检测目标小、背景复杂、计算量大导致的检测精度不足、实时性差等问题,该文提出一种基于YOLOv11n改进的无人机公路日常巡检算法YOLOv11-UR。该算法将MLCA注意力机制加入主干网络,融合通道与空间维度信息,整合局部与全局感受野,有效增强特征表达能力。在仅增加少量参数的前提下,实现了检测精度的显著提升。在Neck部引入GSConv替换标准卷积,使卷积计算的输出尽可能接近标准卷积的同时,降低计算成本;引入VoVGSCSP替换C3k2,降低模型参数量和计算复杂度的同时增强特征提取能力。试验结果表明:YOLOv11-UR算法在无人机公路日常巡检方面具有显著优势,在未损失过多推理速度的情况下,有效减少了模型的参数量与计算开销,且模型检测精确率(Precision,RP)和平均精度均值(Mean Average Precision,mAP)分别达到78.26%和73.34%。改进算法兼顾检测精度与推理效率,能够更好地满足无人机公路日常巡检需求。展开更多
The growing demands on global infrastructure highlight the critical need for durable and efficient pavement systems,particularly under the stress of repetitive heavy traffic loads.The use of geosynthetics within the p...The growing demands on global infrastructure highlight the critical need for durable and efficient pavement systems,particularly under the stress of repetitive heavy traffic loads.The use of geosynthetics within the pavement structure increases the load-carrying capacity of unbound pavement layers by providing lateral restraint,improving vertical stress distribution,and enhancing bearing capacity.Such reinforcement typically aims to either improve the service life of pavements or achieve equivalent performance with a reduced granular cover.Previous and ongoing research quantifies geosynthetic performance in pavement reinforcement using various testing methods.Among these,laboratory model box tests subjected to cyclic loading are pivotal,as they closely replicate real-world traffic conditions.Hence,these studies are essential for understanding how geosynthetics distribute loads and enhance pavement durability.This facilitates the development of optimized geosynthetic design and installation practices,accelerating the loading process to simulate years of traffic wear in a shorter period.This review discusses the improved rutting resistance of unbound pavements reinforced with geosynthetic materials,specifically drawing on data from cyclic plate load tests conducted on laboratory model boxes,as highlighted in the literature.Key variables such as optimum geosynthetic placement,geosynthetic material properties,performance of different geosynthetic materials and the effects of aperture shape and size on rutting resistance are discussed.Furthermore,the review assesses various predictive rutting models,analysing their applicability and accuracy in forecasting the rutting performance of geosynthetic-reinforced unbound pavements.This comprehensive literature review aids pavement engineers and researchers,in guiding the selection and design of geosynthetics to optimize pavement durability and functionality under repetitive traffic loads.展开更多
文摘针对无人机公路日常巡检任务中检测目标小、背景复杂、计算量大导致的检测精度不足、实时性差等问题,该文提出一种基于YOLOv11n改进的无人机公路日常巡检算法YOLOv11-UR。该算法将MLCA注意力机制加入主干网络,融合通道与空间维度信息,整合局部与全局感受野,有效增强特征表达能力。在仅增加少量参数的前提下,实现了检测精度的显著提升。在Neck部引入GSConv替换标准卷积,使卷积计算的输出尽可能接近标准卷积的同时,降低计算成本;引入VoVGSCSP替换C3k2,降低模型参数量和计算复杂度的同时增强特征提取能力。试验结果表明:YOLOv11-UR算法在无人机公路日常巡检方面具有显著优势,在未损失过多推理速度的情况下,有效减少了模型的参数量与计算开销,且模型检测精确率(Precision,RP)和平均精度均值(Mean Average Precision,mAP)分别达到78.26%和73.34%。改进算法兼顾检测精度与推理效率,能够更好地满足无人机公路日常巡检需求。
基金financial and intellectual support provided by Queensland University of Technology(QUT)through its Higher Degree Research Program.
文摘The growing demands on global infrastructure highlight the critical need for durable and efficient pavement systems,particularly under the stress of repetitive heavy traffic loads.The use of geosynthetics within the pavement structure increases the load-carrying capacity of unbound pavement layers by providing lateral restraint,improving vertical stress distribution,and enhancing bearing capacity.Such reinforcement typically aims to either improve the service life of pavements or achieve equivalent performance with a reduced granular cover.Previous and ongoing research quantifies geosynthetic performance in pavement reinforcement using various testing methods.Among these,laboratory model box tests subjected to cyclic loading are pivotal,as they closely replicate real-world traffic conditions.Hence,these studies are essential for understanding how geosynthetics distribute loads and enhance pavement durability.This facilitates the development of optimized geosynthetic design and installation practices,accelerating the loading process to simulate years of traffic wear in a shorter period.This review discusses the improved rutting resistance of unbound pavements reinforced with geosynthetic materials,specifically drawing on data from cyclic plate load tests conducted on laboratory model boxes,as highlighted in the literature.Key variables such as optimum geosynthetic placement,geosynthetic material properties,performance of different geosynthetic materials and the effects of aperture shape and size on rutting resistance are discussed.Furthermore,the review assesses various predictive rutting models,analysing their applicability and accuracy in forecasting the rutting performance of geosynthetic-reinforced unbound pavements.This comprehensive literature review aids pavement engineers and researchers,in guiding the selection and design of geosynthetics to optimize pavement durability and functionality under repetitive traffic loads.
