Consequent on MHD geometry sensibility phenomena was measured in an accident case;the more detail experiments have been conducted at the liquid metal experimental loop upgrade facility (LMEL-U). The experimental resul...Consequent on MHD geometry sensibility phenomena was measured in an accident case;the more detail experiments have been conducted at the liquid metal experimental loop upgrade facility (LMEL-U). The experimental results indicate that MHD pressure drop can be greatly reduced in the special designed ducts. Base on the experimental data, an innovation channel concept (tentatively called as the secondary flow channel, short in “S-channel”) is addressed as a reducing MHD pressure drop channel for the application of a liquid metal blanket system in fusion reactor. It may be a dawn for solving MHD pressure drop key issue of liquid metal blanket system.展开更多
针对乒乓球目标检测方法易受环境、光线、速度等多种因素干扰导致精度和实时性不佳的问题,提出了一种基于YOLOv5s框架的轻量化乒乓球目标检测算法——SYOLO5(Shuffle-YOLOv5s)。首先,采用改进的ShuffleNetV2网络单元组合重构YOLOv5s主...针对乒乓球目标检测方法易受环境、光线、速度等多种因素干扰导致精度和实时性不佳的问题,提出了一种基于YOLOv5s框架的轻量化乒乓球目标检测算法——SYOLO5(Shuffle-YOLOv5s)。首先,采用改进的ShuffleNetV2网络单元组合重构YOLOv5s主干网络,提高特征提取速度;其次,在特征融合的过程中引入高效通道注意力(ECA)机制,有效提升模型的检测性能;接着,采用SIoU Loss(S-Intersection over Union)作为定位损失函数提升网络的收敛速度和定位精度;最后,贴合乒乓球小尺寸的特点,采用双尺度目标检测,进一步提高模型推理速度。实验结果表明,所提算法与YOLOv5s相比,参数量和计算量分别减少了80%和60%,精确率提升了1.9个百分点。展开更多
文摘Consequent on MHD geometry sensibility phenomena was measured in an accident case;the more detail experiments have been conducted at the liquid metal experimental loop upgrade facility (LMEL-U). The experimental results indicate that MHD pressure drop can be greatly reduced in the special designed ducts. Base on the experimental data, an innovation channel concept (tentatively called as the secondary flow channel, short in “S-channel”) is addressed as a reducing MHD pressure drop channel for the application of a liquid metal blanket system in fusion reactor. It may be a dawn for solving MHD pressure drop key issue of liquid metal blanket system.
文摘针对乒乓球目标检测方法易受环境、光线、速度等多种因素干扰导致精度和实时性不佳的问题,提出了一种基于YOLOv5s框架的轻量化乒乓球目标检测算法——SYOLO5(Shuffle-YOLOv5s)。首先,采用改进的ShuffleNetV2网络单元组合重构YOLOv5s主干网络,提高特征提取速度;其次,在特征融合的过程中引入高效通道注意力(ECA)机制,有效提升模型的检测性能;接着,采用SIoU Loss(S-Intersection over Union)作为定位损失函数提升网络的收敛速度和定位精度;最后,贴合乒乓球小尺寸的特点,采用双尺度目标检测,进一步提高模型推理速度。实验结果表明,所提算法与YOLOv5s相比,参数量和计算量分别减少了80%和60%,精确率提升了1.9个百分点。
