针对菌落图像中小菌落易漏检的问题,提出了一种基于INC4-YOLO(you only look once)的计数方法,实现精准的菌落计数。采用带残差结构的Inception模块(Inception module with residual connection,IncRes)替换YOLOv5骨干网络中的Bottlenec...针对菌落图像中小菌落易漏检的问题,提出了一种基于INC4-YOLO(you only look once)的计数方法,实现精准的菌落计数。采用带残差结构的Inception模块(Inception module with residual connection,IncRes)替换YOLOv5骨干网络中的Bottleneck模块,以增强图像特征提取能力。从网络的浅层特征中引出一个小目标检测头,以增强算法在训练过程中对小菌落的注意力。分别在标注微生物自动识别数据集(annotated germs for automated recognition,AGAR)和真实菌落计数场景下对INC4-YOLO进行计数性能测试。实验结果表明,在AGAR测试集中,提出的算法在小菌落的平均百分比绝对值计数误差(mean absolute percentage error,MAPE)比其他先进目标检测算法降低了2%;真实菌落计数场景下,INC4-YOLO的MAPE相比YOLOv5降低了7%,表明该算法可帮助菌落计数设备实现精准计数。展开更多
Security and safety remain paramount concerns for both governments and individuals worldwide.In today’s context,the frequency of crimes and terrorist attacks is alarmingly increasing,becoming increasingly intolerable...Security and safety remain paramount concerns for both governments and individuals worldwide.In today’s context,the frequency of crimes and terrorist attacks is alarmingly increasing,becoming increasingly intolerable to society.Consequently,there is a pressing need for swift identification of potential threats to preemptively alert law enforcement and security forces,thereby preventing potential attacks or violent incidents.Recent advancements in big data analytics and deep learning have significantly enhanced the capabilities of computer vision in object detection,particularly in identifying firearms.This paper introduces a novel automatic firearm detection surveillance system,utilizing a one-stage detection approach named MARIE(Mechanism for Realtime Identification of Firearms).MARIE incorporates the Single Shot Multibox Detector(SSD)model,which has been specifically optimized to balance the speed-accuracy trade-off critical in firearm detection applications.The SSD model was further refined by integrating MobileNetV2 and InceptionV2 architectures for superior feature extraction capabilities.The experimental results demonstrate that this modified SSD configuration provides highly satisfactory performance,surpassing existing methods trained on the same dataset in terms of the critical speedaccuracy trade-off.Through these innovations,MARIE sets a new standard in surveillance technology,offering a robust solution to enhance public safety effectively.展开更多
随着深度学习技术的日益精进,它在植物病害识别领域的应用研究日趋深入,而优化AlexNet模型能有效提升桑叶病害识别的性能。因此,选用AlexNet作为基础网络,针对AlexNet的主干网络和多尺度特征融合策略进行改进,并提出一个新型的网络架构...随着深度学习技术的日益精进,它在植物病害识别领域的应用研究日趋深入,而优化AlexNet模型能有效提升桑叶病害识别的性能。因此,选用AlexNet作为基础网络,针对AlexNet的主干网络和多尺度特征融合策略进行改进,并提出一个新型的网络架构——IP-AlexNet模型。首先,在卷积层之后,引入Inception模块,以捕获桑叶病害图像的多样化特征,并通过减少卷积核降低网络计算的复杂度;其次,利用金字塔卷积进行多尺度特征融合,以增强模型的准确性和鲁棒性;再次,加入SE(Squeeze and Excitation)注意力机制,使模型能够聚焦于图像中的关键区域或特征,从而提高识别的精确度和效率;最后,使用自适应平均池化替换传统的最大池化以生成更平滑的特征图,从而减少图像特征信息的损失。实验结果表明,IP-AlexNet模型在桑叶病害识别方面取得了较好的效果,识别准确率高达95.33%,较AlexNet模型提升了9.66个百分点。另外,精准率、召回率、F1值和混淆矩阵等多元评价指标的综合分析表明,IP-AlexNet模型具有很好的鲁棒性和稳定性。展开更多
本文提出了一种改进U-Net散斑抑制方法,该方法结合了Inception、残差结构和注意力模块,应用于具有不同噪声级别的包裹相位图像。将所提出的方法与传统的降噪方法以及现有的深度学习降噪方法进行了对比,仿真与实验结果表明,所提出的方法...本文提出了一种改进U-Net散斑抑制方法,该方法结合了Inception、残差结构和注意力模块,应用于具有不同噪声级别的包裹相位图像。将所提出的方法与传统的降噪方法以及现有的深度学习降噪方法进行了对比,仿真与实验结果表明,所提出的方法在不同噪声级别下具有更好的散斑抑制效果。此外,我们对降噪后的包裹相位进行了相位重建,对比了不同方法降噪后的相位精度,结果表明,该方法在实际应用中能够有效抑制散斑噪声,取得了较好的效果。This paper proposes an improved U-Net speckle suppression method that integrates Inception and residual structures with attention modules, applied to wrapped phase images with different noise levels. The proposed method is compared with traditional denoising methods as well as existing deep learning-based denoising techniques. Experimental results show that our method achieves better speckle suppression across various noise levels. Furthermore, we performed phase reconstruction on the denoised wrapped phase images and compared the phase accuracy of different denoising methods. The results show that the proposed method can effectively suppress speckle noise in practical applications and achieve satisfactory performance.展开更多
针对锂电池健康状态(State of Health,SOH)估计和剩余使用寿命(Remaining Useful Life,RUL)预测过程中健康特征提取单一、估计精度低等问题,提出了一种Inception-LSTM模型用于锂电池SOH估计与RUL预测。首先选取合适的恒压恒流充电时间...针对锂电池健康状态(State of Health,SOH)估计和剩余使用寿命(Remaining Useful Life,RUL)预测过程中健康特征提取单一、估计精度低等问题,提出了一种Inception-LSTM模型用于锂电池SOH估计与RUL预测。首先选取合适的恒压恒流充电时间构建特征序列HF,并采用Pearson相关性系数分析HF和容量之间的相关性;另外针对特征变量的特征提取不够全面问题,采用Inception模型进行特征提取,采用LSTM进行时序建模,随后利用注意力机制进一步提取对电池健康度影响较大的特征来估计电池健康状态,利用该深度学习模型来挖掘电池在复杂使用条件下的动态变化特征。实验结果表明文章模型SOH估计最大均方根误差在3.86%以内,RUL预测最大误差在1个循环。实验结果表明该方法在SOH估计和RUL预测方面优于传统模型。展开更多
文摘针对菌落图像中小菌落易漏检的问题,提出了一种基于INC4-YOLO(you only look once)的计数方法,实现精准的菌落计数。采用带残差结构的Inception模块(Inception module with residual connection,IncRes)替换YOLOv5骨干网络中的Bottleneck模块,以增强图像特征提取能力。从网络的浅层特征中引出一个小目标检测头,以增强算法在训练过程中对小菌落的注意力。分别在标注微生物自动识别数据集(annotated germs for automated recognition,AGAR)和真实菌落计数场景下对INC4-YOLO进行计数性能测试。实验结果表明,在AGAR测试集中,提出的算法在小菌落的平均百分比绝对值计数误差(mean absolute percentage error,MAPE)比其他先进目标检测算法降低了2%;真实菌落计数场景下,INC4-YOLO的MAPE相比YOLOv5降低了7%,表明该算法可帮助菌落计数设备实现精准计数。
文摘Security and safety remain paramount concerns for both governments and individuals worldwide.In today’s context,the frequency of crimes and terrorist attacks is alarmingly increasing,becoming increasingly intolerable to society.Consequently,there is a pressing need for swift identification of potential threats to preemptively alert law enforcement and security forces,thereby preventing potential attacks or violent incidents.Recent advancements in big data analytics and deep learning have significantly enhanced the capabilities of computer vision in object detection,particularly in identifying firearms.This paper introduces a novel automatic firearm detection surveillance system,utilizing a one-stage detection approach named MARIE(Mechanism for Realtime Identification of Firearms).MARIE incorporates the Single Shot Multibox Detector(SSD)model,which has been specifically optimized to balance the speed-accuracy trade-off critical in firearm detection applications.The SSD model was further refined by integrating MobileNetV2 and InceptionV2 architectures for superior feature extraction capabilities.The experimental results demonstrate that this modified SSD configuration provides highly satisfactory performance,surpassing existing methods trained on the same dataset in terms of the critical speedaccuracy trade-off.Through these innovations,MARIE sets a new standard in surveillance technology,offering a robust solution to enhance public safety effectively.
文摘随着深度学习技术的日益精进,它在植物病害识别领域的应用研究日趋深入,而优化AlexNet模型能有效提升桑叶病害识别的性能。因此,选用AlexNet作为基础网络,针对AlexNet的主干网络和多尺度特征融合策略进行改进,并提出一个新型的网络架构——IP-AlexNet模型。首先,在卷积层之后,引入Inception模块,以捕获桑叶病害图像的多样化特征,并通过减少卷积核降低网络计算的复杂度;其次,利用金字塔卷积进行多尺度特征融合,以增强模型的准确性和鲁棒性;再次,加入SE(Squeeze and Excitation)注意力机制,使模型能够聚焦于图像中的关键区域或特征,从而提高识别的精确度和效率;最后,使用自适应平均池化替换传统的最大池化以生成更平滑的特征图,从而减少图像特征信息的损失。实验结果表明,IP-AlexNet模型在桑叶病害识别方面取得了较好的效果,识别准确率高达95.33%,较AlexNet模型提升了9.66个百分点。另外,精准率、召回率、F1值和混淆矩阵等多元评价指标的综合分析表明,IP-AlexNet模型具有很好的鲁棒性和稳定性。
文摘本文提出了一种改进U-Net散斑抑制方法,该方法结合了Inception、残差结构和注意力模块,应用于具有不同噪声级别的包裹相位图像。将所提出的方法与传统的降噪方法以及现有的深度学习降噪方法进行了对比,仿真与实验结果表明,所提出的方法在不同噪声级别下具有更好的散斑抑制效果。此外,我们对降噪后的包裹相位进行了相位重建,对比了不同方法降噪后的相位精度,结果表明,该方法在实际应用中能够有效抑制散斑噪声,取得了较好的效果。This paper proposes an improved U-Net speckle suppression method that integrates Inception and residual structures with attention modules, applied to wrapped phase images with different noise levels. The proposed method is compared with traditional denoising methods as well as existing deep learning-based denoising techniques. Experimental results show that our method achieves better speckle suppression across various noise levels. Furthermore, we performed phase reconstruction on the denoised wrapped phase images and compared the phase accuracy of different denoising methods. The results show that the proposed method can effectively suppress speckle noise in practical applications and achieve satisfactory performance.
文摘针对锂电池健康状态(State of Health,SOH)估计和剩余使用寿命(Remaining Useful Life,RUL)预测过程中健康特征提取单一、估计精度低等问题,提出了一种Inception-LSTM模型用于锂电池SOH估计与RUL预测。首先选取合适的恒压恒流充电时间构建特征序列HF,并采用Pearson相关性系数分析HF和容量之间的相关性;另外针对特征变量的特征提取不够全面问题,采用Inception模型进行特征提取,采用LSTM进行时序建模,随后利用注意力机制进一步提取对电池健康度影响较大的特征来估计电池健康状态,利用该深度学习模型来挖掘电池在复杂使用条件下的动态变化特征。实验结果表明文章模型SOH估计最大均方根误差在3.86%以内,RUL预测最大误差在1个循环。实验结果表明该方法在SOH估计和RUL预测方面优于传统模型。
