To address the problem that it is difficult to detect an intermediate frequency(IF)signal at the receiving end of a communication system under extremely low signal-to-noise ratio(SNR)conditions,we propose a stochastic...To address the problem that it is difficult to detect an intermediate frequency(IF)signal at the receiving end of a communication system under extremely low signal-to-noise ratio(SNR)conditions,we propose a stochastic resonance(SR)-enhanced sine-signal detection method based on the sign function.By analyzing the SR mechanism of the sine signal and combining it with the characteristics of a dual-sequence frequency-hopping(DSFH)receiver,a periodic stationary solution of the Fokker-Planck equation(FPE)with a time parameter is obtained.The extreme point of the sine signal is selected as the decision time,and the force law of the electromagnetic particles is analyzed.A receiving structure based on the sign function is proposed to maximize the output difference of the system,and the value condition of the sign function is determined.In order to further improve the detection performance,in combination with the central-limit theorem,the sampling points are averaged N times,and the signal-detection problem is transformed into a hypothesis-testing problem under a Gaussian distribution.The theoretical analysis and simulation experiment results confirm that when N is 100 and the SNR is greater than 20 dB,the bit-error ratio(BER)is less than 1.5×10^(-2) under conditions in which the signal conforms to the optimal SR parameters.展开更多
针对复杂道路场景下现有算法对道路上交通标志检测精度不高以及漏检误检问题,文中提出了一种改进YOLOv5(You Only Look Once version 5)检测算法。在主干网络使用轻量化网络RepVGG(Re-parameterized Visual Geometry Goup)增强模型特征...针对复杂道路场景下现有算法对道路上交通标志检测精度不高以及漏检误检问题,文中提出了一种改进YOLOv5(You Only Look Once version 5)检测算法。在主干网络使用轻量化网络RepVGG(Re-parameterized Visual Geometry Goup)增强模型特征提取能力,通过增加残差分支结构获得更多特征细节信息。将坐标注意力机制(Coordinate Attention,CA)融入RepVGG模块来提升模型对于小目标的识别与定位能力。在特征融合方面,引入加权双向特征金字塔结构(Bi-directional Feature Pyramid Network,BiFPN)重构颈部网络,充分利用不同尺度特征信息来增强网络特征融合能力。将原CIoU(Complete Intersection over Union)损失函数替换为EIoU(Efficient IoU)损失函数,以此提高回归框的稳定性,加速收敛。实验结果表明,改进算法在CCTSDB(Chinese Traffic Sign Detection Bench mark)数据集的均值平均精度为98.9%。相较于原YOLOv5算法,所提算法的平均精度提升了2.5百分点,召回率提升了4.5百分点,减少了漏检和误检发生概率,同时满足实时检测要求。展开更多
现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有...现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有像素级建模能力,提高它对小目标交通标志特征的提取能力,YOLOv7-NCAM算法使用FReLU激活函数构建了DBF和CBF两种卷积层,并用它们来组建模型的Backbone模块和Neck模块;提出一种归一化通道注意力机制(normalized channel attention mechanism,NCAM)并加入Head模块中。通过与整体网络一起训练,得到归一化(batch normalization,BN)缩放因子,利用缩放因子算出各个通道的权重因子,提升网络对交通标志特征的表达能力,从而使YOLOv7-NCAM网络模型能够集中关注检测目标交通标志。通过在CCTSDB-2021交通标志检测数据集上的测试,与YOLOv7网络模型对比结果表明,YOLOv7-NCAM算法对背景复杂下小交通标志的检测各项指标均有明显提高:准确率(precision,P)达到91.5%,比原网络高出9.5个百分点;召回率(recall,R)达到85.9%,比原网络高出5.7个百分点;均值平均精度(mean average precision,mAP)达到了91.4%,比原网络高出4.7个百分点。与现有的交通标志检测算法相比,YOLOv7-NCAM算法的检测准确率也有提高,且检测速度48.3 FPS,能满足实时需求。展开更多
基金Project supported by the Natural Science Foundation of Hebei Province of China(Grant Nos.F2019506031,F2019506037,and F2020506036)the Frontier Innovation Program of Army Engineering University(Grant No.KYSZJQZL2005)the Basic Frontier Science and Technology Innovation Program of Army Engineering University(Grant No.KYSZJQZL2020).
文摘To address the problem that it is difficult to detect an intermediate frequency(IF)signal at the receiving end of a communication system under extremely low signal-to-noise ratio(SNR)conditions,we propose a stochastic resonance(SR)-enhanced sine-signal detection method based on the sign function.By analyzing the SR mechanism of the sine signal and combining it with the characteristics of a dual-sequence frequency-hopping(DSFH)receiver,a periodic stationary solution of the Fokker-Planck equation(FPE)with a time parameter is obtained.The extreme point of the sine signal is selected as the decision time,and the force law of the electromagnetic particles is analyzed.A receiving structure based on the sign function is proposed to maximize the output difference of the system,and the value condition of the sign function is determined.In order to further improve the detection performance,in combination with the central-limit theorem,the sampling points are averaged N times,and the signal-detection problem is transformed into a hypothesis-testing problem under a Gaussian distribution.The theoretical analysis and simulation experiment results confirm that when N is 100 and the SNR is greater than 20 dB,the bit-error ratio(BER)is less than 1.5×10^(-2) under conditions in which the signal conforms to the optimal SR parameters.
文摘现有目标检测算法对背景复杂下小交通标志的检测效果并不理想。为此,提出了一种基于归一化通道注意力机制YOLOv7的交通标志检测算法(YOLOv7 based on normalized channel attention mechanism,YOLOv7-NCAM)。为了使YOLOv7-NCAM模型具有像素级建模能力,提高它对小目标交通标志特征的提取能力,YOLOv7-NCAM算法使用FReLU激活函数构建了DBF和CBF两种卷积层,并用它们来组建模型的Backbone模块和Neck模块;提出一种归一化通道注意力机制(normalized channel attention mechanism,NCAM)并加入Head模块中。通过与整体网络一起训练,得到归一化(batch normalization,BN)缩放因子,利用缩放因子算出各个通道的权重因子,提升网络对交通标志特征的表达能力,从而使YOLOv7-NCAM网络模型能够集中关注检测目标交通标志。通过在CCTSDB-2021交通标志检测数据集上的测试,与YOLOv7网络模型对比结果表明,YOLOv7-NCAM算法对背景复杂下小交通标志的检测各项指标均有明显提高:准确率(precision,P)达到91.5%,比原网络高出9.5个百分点;召回率(recall,R)达到85.9%,比原网络高出5.7个百分点;均值平均精度(mean average precision,mAP)达到了91.4%,比原网络高出4.7个百分点。与现有的交通标志检测算法相比,YOLOv7-NCAM算法的检测准确率也有提高,且检测速度48.3 FPS,能满足实时需求。
