面向大规模感知与智能应用场景,集中式计算在时延、带宽、能耗与隐私保护的多重约束下逐渐呈现边际效益递减,计算范式因此由单一的“万物上云”模式,逐步转向“就地计算与云边协同”的新形态。在此背景下,本文首先梳理集中化计算路径在...面向大规模感知与智能应用场景,集中式计算在时延、带宽、能耗与隐私保护的多重约束下逐渐呈现边际效益递减,计算范式因此由单一的“万物上云”模式,逐步转向“就地计算与云边协同”的新形态。在此背景下,本文首先梳理集中化计算路径在不同发展阶段所具备的优势及其适用边界,进而界定边缘计算在端-云之间所扮演的关键角色。在此基础上,进一步概述“传感云-边缘-端”协同计算框架,重点分析其中的核心机制,包括数据“必要即上行”的传输原则、面向服务级别协议(SLA)感知的任务分配与双层调度策略,以及边侧即时闭环执行与云侧全局策略治理之间的分工与协同关系。随着计算与智能能力向边缘侧持续下沉,本文进一步讨论边缘智能的发展方向,涵盖模型轻量化与本地学习机制、联邦学习与知识蒸馏的协同范式,以及面向边缘环境的智能运维(AIOps for Edge)与多级降级机制所支撑的自治能力。同时,强调构建以端到端闭环效率、系统韧性与可追责性为导向的综合评价体系的重要性。最后,结合教育等典型应用场景以及产业实践,论证就地计算与云边协同在保障确定性时延、提升系统整体韧性以及实现跨域一致性方面的现实有效性,并据此指出计算范式由边缘计算向云边智能协同演进的必然趋势与发展方向。展开更多
This paper describes a 2D/3D vision chip with integrated sensing and processing capabilities.The 2D/3D vision chip architecture includes a 2D/3D image sensor and a programmable visual processor.In this architecture,we...This paper describes a 2D/3D vision chip with integrated sensing and processing capabilities.The 2D/3D vision chip architecture includes a 2D/3D image sensor and a programmable visual processor.In this architecture,we design a novel on-chip processing flow with die-to-die image transmission and low-latency fixed-point image processing.The vision chip achieves real-time end-to-end processing of convolutional neural networks(CNNs)and conventional image processing algo-rithms.Furthermore,an end-to-end 2D/3D vision system is built to exhibit the capacity of the vision chip.The vision system achieves real-timing applications under 2D and 3D scenes,such as human face detection(processing delay 10.2 ms)and depth map reconstruction(processing delay 4.1 ms).The frame rate of image acquisition,image process,and result display is larger than 30 fps.展开更多
Ten physical and environmental variables collected from an on-the-go soil sensor at two field sites (MF3E and MF11S) in Mississippi, USA, were analyzed to assess soil variability and the interrelationships among the m...Ten physical and environmental variables collected from an on-the-go soil sensor at two field sites (MF3E and MF11S) in Mississippi, USA, were analyzed to assess soil variability and the interrelationships among the measurements. At MF3E, moderate variability was observed in apparent electrical conductivity shallow (ECas), slope, and ECa ratio measurements, with coefficients of variation ranging from 20% to 27%. In contrast, MF11S exhibited higher variability, particularly in ECas and ECad (deep) measurements, which exceeded 30% in their coefficient of variation values, indicating significant differences in soil composition and moisture content. Correlation analysis revealed strong positive relationships between the near-infrared-to-red ratio and red reflectance (r = 0.897***) soil values at MF3E. MF11S demonstrated a strong negative correlation between ECas and ECad readings with the x-coordinate (r ***). Scatter plots and fitted models illustrated the complexity of relationships, with many showing nonlinear trends. These findings emphasize the need for continuous monitoring and advanced modeling to understand the dynamic nature of soil properties and their implications for agricultural practices. Future research should explore the underlying mechanisms driving variability in the soil characteristics to enhance soil management strategies at the study sites.展开更多
The latest developments in bio-inspired neuromorphic vision sensors can be summarized in 3 keywords:smaller,faster,and smarter.(1)Smaller:Devices are becoming more compact by integrating previously separated component...The latest developments in bio-inspired neuromorphic vision sensors can be summarized in 3 keywords:smaller,faster,and smarter.(1)Smaller:Devices are becoming more compact by integrating previously separated components such as sensors,memory,and processing units.As a prime example,the transition from traditional sensory vision computing to in-sensor vision computing has shown clear benefits,such as simpler circuitry,lower power consumption,and less data redundancy.(2)Swifter:Owing to the nature of physics,smaller and more integrated devices can detect,process,and react to input more quickly.In addition,the methods for sensing and processing optical information using various materials(such as oxide semiconductors)are evolving.(3)Smarter:Owing to these two main research directions,we can expect advanced applications such as adaptive vision sensors,collision sensors,and nociceptive sensors.This review mainly focuses on the recent progress,working mechanisms,image pre-processing techniques,and advanced features of two types of neuromorphic vision sensors based on near-sensor and in-sensor vision computing methodologies.展开更多
洪涝过程探测与准实时服务对于保障人民生命和财产安全意义重大。通常洪涝探测仅关注洪涝断面/平均状态,缺乏对洪涝从发生、发展乃至最后结束的全过程的整体认知,且洪涝探测和服务被动滞后。本文制定洪涝过程探测规则,改进水位预测模型...洪涝过程探测与准实时服务对于保障人民生命和财产安全意义重大。通常洪涝探测仅关注洪涝断面/平均状态,缺乏对洪涝从发生、发展乃至最后结束的全过程的整体认知,且洪涝探测和服务被动滞后。本文制定洪涝过程探测规则,改进水位预测模型,并将其作为理论基础,与传感网信息模型和服务接口相结合,提出了洪涝过程动态探测与准实时服务(process-based flood detection and service,PFD&S)方法。以PFD&S方法为基础,设计并开发了PFD&S原型系统,该系统由传感器层、数据接入层、洪涝探测层和用户交互层4层组成,且具备传感器数据发布和洪涝事件订阅两种使用模式。文中以2016年夏季发生在黄汉流域的洪涝事件为例,对PFD&S方法的可行性和有效性进行验证。结果表明,PFD&S方法和系统能够精准判断洪涝事件所处阶段,并根据不同阶段的需求提供水位预测、洪涝预警或洪涝信息统计等服务,且方法具备准实时性和可扩展性特征。展开更多
文摘面向大规模感知与智能应用场景,集中式计算在时延、带宽、能耗与隐私保护的多重约束下逐渐呈现边际效益递减,计算范式因此由单一的“万物上云”模式,逐步转向“就地计算与云边协同”的新形态。在此背景下,本文首先梳理集中化计算路径在不同发展阶段所具备的优势及其适用边界,进而界定边缘计算在端-云之间所扮演的关键角色。在此基础上,进一步概述“传感云-边缘-端”协同计算框架,重点分析其中的核心机制,包括数据“必要即上行”的传输原则、面向服务级别协议(SLA)感知的任务分配与双层调度策略,以及边侧即时闭环执行与云侧全局策略治理之间的分工与协同关系。随着计算与智能能力向边缘侧持续下沉,本文进一步讨论边缘智能的发展方向,涵盖模型轻量化与本地学习机制、联邦学习与知识蒸馏的协同范式,以及面向边缘环境的智能运维(AIOps for Edge)与多级降级机制所支撑的自治能力。同时,强调构建以端到端闭环效率、系统韧性与可追责性为导向的综合评价体系的重要性。最后,结合教育等典型应用场景以及产业实践,论证就地计算与云边协同在保障确定性时延、提升系统整体韧性以及实现跨域一致性方面的现实有效性,并据此指出计算范式由边缘计算向云边智能协同演进的必然趋势与发展方向。
基金supported in part by the National Key Research and Development Program of China(Grant No.2019YFB2204300)in part by the National Natural Science Foundation of China(Grant Nos.62334008 and 62274154)in part by the Key Program of National Natural Science Foundation of China(Grant No.62134004).
文摘This paper describes a 2D/3D vision chip with integrated sensing and processing capabilities.The 2D/3D vision chip architecture includes a 2D/3D image sensor and a programmable visual processor.In this architecture,we design a novel on-chip processing flow with die-to-die image transmission and low-latency fixed-point image processing.The vision chip achieves real-time end-to-end processing of convolutional neural networks(CNNs)and conventional image processing algo-rithms.Furthermore,an end-to-end 2D/3D vision system is built to exhibit the capacity of the vision chip.The vision system achieves real-timing applications under 2D and 3D scenes,such as human face detection(processing delay 10.2 ms)and depth map reconstruction(processing delay 4.1 ms).The frame rate of image acquisition,image process,and result display is larger than 30 fps.
文摘Ten physical and environmental variables collected from an on-the-go soil sensor at two field sites (MF3E and MF11S) in Mississippi, USA, were analyzed to assess soil variability and the interrelationships among the measurements. At MF3E, moderate variability was observed in apparent electrical conductivity shallow (ECas), slope, and ECa ratio measurements, with coefficients of variation ranging from 20% to 27%. In contrast, MF11S exhibited higher variability, particularly in ECas and ECad (deep) measurements, which exceeded 30% in their coefficient of variation values, indicating significant differences in soil composition and moisture content. Correlation analysis revealed strong positive relationships between the near-infrared-to-red ratio and red reflectance (r = 0.897***) soil values at MF3E. MF11S demonstrated a strong negative correlation between ECas and ECad readings with the x-coordinate (r ***). Scatter plots and fitted models illustrated the complexity of relationships, with many showing nonlinear trends. These findings emphasize the need for continuous monitoring and advanced modeling to understand the dynamic nature of soil properties and their implications for agricultural practices. Future research should explore the underlying mechanisms driving variability in the soil characteristics to enhance soil management strategies at the study sites.
基金This work was supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.NRF-2019R1A2C2002447)This research also was supported by Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(No.NRF-2014R1A6A1030419)This work also was supported by Korea Institute for Advancement of Technology(KIAT)grant funded by the Korea Government(MOTIE)(P0020967,Advanced Training Program for Smart Sensor Engineers).
文摘The latest developments in bio-inspired neuromorphic vision sensors can be summarized in 3 keywords:smaller,faster,and smarter.(1)Smaller:Devices are becoming more compact by integrating previously separated components such as sensors,memory,and processing units.As a prime example,the transition from traditional sensory vision computing to in-sensor vision computing has shown clear benefits,such as simpler circuitry,lower power consumption,and less data redundancy.(2)Swifter:Owing to the nature of physics,smaller and more integrated devices can detect,process,and react to input more quickly.In addition,the methods for sensing and processing optical information using various materials(such as oxide semiconductors)are evolving.(3)Smarter:Owing to these two main research directions,we can expect advanced applications such as adaptive vision sensors,collision sensors,and nociceptive sensors.This review mainly focuses on the recent progress,working mechanisms,image pre-processing techniques,and advanced features of two types of neuromorphic vision sensors based on near-sensor and in-sensor vision computing methodologies.
文摘洪涝过程探测与准实时服务对于保障人民生命和财产安全意义重大。通常洪涝探测仅关注洪涝断面/平均状态,缺乏对洪涝从发生、发展乃至最后结束的全过程的整体认知,且洪涝探测和服务被动滞后。本文制定洪涝过程探测规则,改进水位预测模型,并将其作为理论基础,与传感网信息模型和服务接口相结合,提出了洪涝过程动态探测与准实时服务(process-based flood detection and service,PFD&S)方法。以PFD&S方法为基础,设计并开发了PFD&S原型系统,该系统由传感器层、数据接入层、洪涝探测层和用户交互层4层组成,且具备传感器数据发布和洪涝事件订阅两种使用模式。文中以2016年夏季发生在黄汉流域的洪涝事件为例,对PFD&S方法的可行性和有效性进行验证。结果表明,PFD&S方法和系统能够精准判断洪涝事件所处阶段,并根据不同阶段的需求提供水位预测、洪涝预警或洪涝信息统计等服务,且方法具备准实时性和可扩展性特征。
