在交通标志检测任务中,交通标志的准确识别对于自动驾驶和智能交通系统至关重要。本文对Yolov8s模型改进并在CCTSDB2021数据集中进行实验以评估模型的性能,我们采取了以下改进,引入了专门针对小目标设计的检测头,该检测头通过优化特征...在交通标志检测任务中,交通标志的准确识别对于自动驾驶和智能交通系统至关重要。本文对Yolov8s模型改进并在CCTSDB2021数据集中进行实验以评估模型的性能,我们采取了以下改进,引入了专门针对小目标设计的检测头,该检测头通过优化特征图的尺度,增强了模型对小尺寸交通标志的识别能力。采用了SIoU损失函数,有助于提升小目标的定位精度。使用模块Star Block集成了一个轻量化模块LB,该模块通过减少参数量同时保持模型检测能力。在CCTSDB2021数据集上的实验结果表明,经过以上改进的Yolov8s模型在检测小尺寸交通标志时,不仅提高了检测准确率,而且减少了一定的参数量,展现了模型改进的有效性。In the traffic sign detection task, accurate recognition of traffic signs is crucial for autonomous driving and intelligent transportation systems. In this paper, we improve the Yolov8s model and conduct experiments in the CCTSDB2021 dataset to evaluate the performance of the model. We have made the following improvements: introducing a detection head specifically designed for small targets. This detection head enhances the model’s recognition ability for small-sized traffic signs by optimizing the scale of the feature map. The SIoU loss function is adopted, which helps to improve the positioning accuracy of small targets. Using the module Star Block, we integrated a lightweight module LB. This module reduces the number of parameters while maintaining the model’s detection ability. The experimental results on the CCTSDB2021 dataset show that the improved Yolov8s model not only improves the detection accuracy but also reduces a certain number of parameters when detecting small-sized traffic signs, demonstrating the effectiveness of model improvement.展开更多
The novel BaTiO3/BiFeO3/TiO2 multilayer heterojunction is prepared on a fluorine-doped tinoxide(FTO) substrate by the sol–gel method. The results indicate that the Pt/Ba TiO3/BiFeO3/TiO2/FTO heterojunction exhibits s...The novel BaTiO3/BiFeO3/TiO2 multilayer heterojunction is prepared on a fluorine-doped tinoxide(FTO) substrate by the sol–gel method. The results indicate that the Pt/Ba TiO3/BiFeO3/TiO2/FTO heterojunction exhibits stable bipolar resistive switching characteristic, good retention performance, and reversal characteristic. Under different pulse voltages and light fields, four stable resistance states can also be realized. The analysis shows that the main conduction mechanism of the resistive switching characteristic of the heterojunction is space charge limited current(SCLC) effect. After the comprehensive analysis of the band diagram and the P–E ferroelectric property of the multilayer heterojunction, we can deduce that the SCLC is formed by the effect of the oxygen vacancy which is controlled by ferroelectric polarizationmodulated change of interfacial barrier. And the effective photo-generated carrier also plays a regulatory role in resistance state(RS), which is formed by the double ferroelectric layer Ba TiO3/BiFeO3 under different light fields. This research is of potential application values for developing the multi-state non-volatile resistance random access memory(RRAM) devices based on ferroelectric materials.展开更多
In this paper,we present a fast mode decomposition method for few-mode fibers,utilizing a lightweight neural network called MobileNetV3-Light.This method can quickly and accurately predict the amplitude and phase info...In this paper,we present a fast mode decomposition method for few-mode fibers,utilizing a lightweight neural network called MobileNetV3-Light.This method can quickly and accurately predict the amplitude and phase information of different modes,enabling us to fully characterize the optical field without the need for expensive experimental equipment.We train the MobileNetV3-Light using simulated near-field optical field maps,and evaluate its performance using both simulated and reconstructed near-field optical field maps.To validate the effectiveness of this method,we conduct mode decomposition experiments on a few-mode fiber supporting six linear polarization(LP)modes(LP01,LP11e,LP11o,LP21e,LP21o,LP02).The results demonstrate a remarkable average correlation of 0.9995 between our simulated and reconstructed near-field lightfield maps.And the mode decomposition speed is about 6 ms per frame,indicating its powerful real-time processing capability.In addition,the proposed network model is compact,with a size of only 6.5 MB,making it well suited for deployment on portable mobile devices.展开更多
文摘在交通标志检测任务中,交通标志的准确识别对于自动驾驶和智能交通系统至关重要。本文对Yolov8s模型改进并在CCTSDB2021数据集中进行实验以评估模型的性能,我们采取了以下改进,引入了专门针对小目标设计的检测头,该检测头通过优化特征图的尺度,增强了模型对小尺寸交通标志的识别能力。采用了SIoU损失函数,有助于提升小目标的定位精度。使用模块Star Block集成了一个轻量化模块LB,该模块通过减少参数量同时保持模型检测能力。在CCTSDB2021数据集上的实验结果表明,经过以上改进的Yolov8s模型在检测小尺寸交通标志时,不仅提高了检测准确率,而且减少了一定的参数量,展现了模型改进的有效性。In the traffic sign detection task, accurate recognition of traffic signs is crucial for autonomous driving and intelligent transportation systems. In this paper, we improve the Yolov8s model and conduct experiments in the CCTSDB2021 dataset to evaluate the performance of the model. We have made the following improvements: introducing a detection head specifically designed for small targets. This detection head enhances the model’s recognition ability for small-sized traffic signs by optimizing the scale of the feature map. The SIoU loss function is adopted, which helps to improve the positioning accuracy of small targets. Using the module Star Block, we integrated a lightweight module LB. This module reduces the number of parameters while maintaining the model’s detection ability. The experimental results on the CCTSDB2021 dataset show that the improved Yolov8s model not only improves the detection accuracy but also reduces a certain number of parameters when detecting small-sized traffic signs, demonstrating the effectiveness of model improvement.
基金Project supported by the Scientific Research Program of Hunan Provincial Education Department,China(Grant No.18C0232)the International Cooperative Extension Program of Changsha University of Science and Technology,China(Grant No.2019IC35)
文摘The novel BaTiO3/BiFeO3/TiO2 multilayer heterojunction is prepared on a fluorine-doped tinoxide(FTO) substrate by the sol–gel method. The results indicate that the Pt/Ba TiO3/BiFeO3/TiO2/FTO heterojunction exhibits stable bipolar resistive switching characteristic, good retention performance, and reversal characteristic. Under different pulse voltages and light fields, four stable resistance states can also be realized. The analysis shows that the main conduction mechanism of the resistive switching characteristic of the heterojunction is space charge limited current(SCLC) effect. After the comprehensive analysis of the band diagram and the P–E ferroelectric property of the multilayer heterojunction, we can deduce that the SCLC is formed by the effect of the oxygen vacancy which is controlled by ferroelectric polarizationmodulated change of interfacial barrier. And the effective photo-generated carrier also plays a regulatory role in resistance state(RS), which is formed by the double ferroelectric layer Ba TiO3/BiFeO3 under different light fields. This research is of potential application values for developing the multi-state non-volatile resistance random access memory(RRAM) devices based on ferroelectric materials.
基金supported by the Scientific Research Fund of Hunan Provincial Education Department of China(No.22B0324)the Natural Science Foundation of Hunan Province of China(No.2020JJ5606)。
文摘In this paper,we present a fast mode decomposition method for few-mode fibers,utilizing a lightweight neural network called MobileNetV3-Light.This method can quickly and accurately predict the amplitude and phase information of different modes,enabling us to fully characterize the optical field without the need for expensive experimental equipment.We train the MobileNetV3-Light using simulated near-field optical field maps,and evaluate its performance using both simulated and reconstructed near-field optical field maps.To validate the effectiveness of this method,we conduct mode decomposition experiments on a few-mode fiber supporting six linear polarization(LP)modes(LP01,LP11e,LP11o,LP21e,LP21o,LP02).The results demonstrate a remarkable average correlation of 0.9995 between our simulated and reconstructed near-field lightfield maps.And the mode decomposition speed is about 6 ms per frame,indicating its powerful real-time processing capability.In addition,the proposed network model is compact,with a size of only 6.5 MB,making it well suited for deployment on portable mobile devices.
