- 标题
- 摘要
- 关键词
- 实验方案
- 产品
过滤筛选
- 2018
- Conditional Random Fields (CRF)
- Convolutional Neural Network (CNN)
- Fine Classification
- Airborne hyperspectral
- green tide
- Elegant End-to-End Fully Convolutional Network (E3FCN)
- deep learning
- remote sensing
- Moderate Resolution Imaging Spectroradiometer (MODIS)
- Optoelectronic Information Science and Engineering
- Ocean University of China
- Wuhan University
- Central South University
- Hubei University
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Road Segmentation of UAV RS Image Using Adversarial Network with Multi-Scale Context Aggregation
摘要： Semantic segmentation using adversarial networks has been approved to produce the better artificial results in image processing fields. Focused on current Deep Convolutional Neural Networks (DCNNs), since the convolutional kernel size has been fixed in every convolutional operation, the small objects would be ignored with large convolutional kernel size, and the segmentation result of large objects is not continuous with small convolutional kernel size. The paper developed a semantic segmentation model that combined the adversarial networks with multi-scale context aggregation. Further, the model was applied to road segmentation of UAV RS images. The experimental results of this semantic segmentation model with multi-scale context aggregation has a better performance for road segmentation and fit well with the reference standard results. It can improve the road segmentation accuracy obviously in the situation where there are other small regions whose shape or color is similar to road regions in UAV RS images.
关键词： Road Segmentation,Adversarial Network,UAV image,Image processing,multi-scale context aggregation
更新于2025-09-23 15:22:29
-
Thermal Mapping of a Lithium Polymer Batteries Pack with FBGs Network
摘要： In this paper, a network of 37 fiber Bragg grating (FBG) sensors is proposed for real-time, in situ, and operando multipoint monitoring of the surface temperature distribution on a pack of three prismatic lithium polymer batteries (LiPBs). Using the network, a spatial and temporal thermal mapping of all pack interfaces was performed. In each interface, nine strategic locations were monitored by considering a three-by-three matrix, corresponding to the LiPBs top, middle and bottom zones. The batteries were subjected to charge and discharge cycles, where the charge was carried out at 1.0 C, whereas the discharge rates were 0.7 C and 1.4 C. The results show that in general, a thermal gradient is recognized from the top to the bottom, but is less prominent in the end-of-charge steps. The results also indicate the presence of hot spots between two of the three batteries, which were located near the positive tab collector. This occurs due to the higher current density of the lithium ions in this area. The presented FBG sensing network can be used to improve the thermal management of batteries by performing a spatiotemporal thermal mapping, as well as by identifying the zones which are more conducive to the possibility of the existence of hot spots, thereby preventing severe consequences such as thermal runaway and promoting their safety. To our knowledge, this is the first time that a spatial and temporal thermal mapping is reported for this specific application using a network of FBG sensors.
关键词： in situ monitoring,lithium polymer batteries pack,FBGs network,safety,thermal mapping
更新于2025-09-23 15:22:29
-
Mechanically Enhanced Electrical Conductivity of Polydimethylsiloxane-Based Composites by a Hot Embossing Process
摘要： Electrically conductive polymer composites are in high demand for modern technologies, however, the intrinsic brittleness of conducting conjugated polymers and the moderate electrical conductivity of engineering polymer/carbon composites have highly constrained their applications. In this work, super high electrical conductive polymer composites were produced by a novel hot embossing design. The polydimethylsiloxane (PDMS) composites containing short carbon fiber (SCF) exhibited an electrical percolation threshold at 0.45 wt % and reached a saturated electrical conductivity of 49 S/m at 8 wt % of SCF. When reducing the sample thickness from 1.0 to 0.1 mm by the hot embossing process, a compression-induced percolation threshold occurred at 0.3 wt %, while the electrical conductivity was further enhanced to 378 S/m at 8 wt % SCF. Furthermore, the addition of a second nanofiller of 1 wt %, such as carbon nanotube or conducting carbon black, further increased the electrical conductivity of the PDMS/SCF (8 wt %) composites to 909 S/m and 657 S/m, respectively. The synergy of the densified conducting filler network by the mechanical compression and the hierarchical micro-/nano-scale filler approach has realized super high electrically conductive, yet mechanically flexible, polymer composites for modern flexible electronics applications.
关键词： compression-induced percolation threshold,synergy,electrical conducting network,hybrid filler,forced assembly
更新于2025-09-23 15:22:29
-
Application research of image recognition technology based on CNN in image location of environmental monitoring UAV
摘要： UAV remote sensing has been widely used in emergency rescue, disaster relief, environmental monitoring, urban planning, and so on. Image recognition and image location in environmental monitoring has become an academic hotspot in the field of computer vision. Convolution neural network model is the most commonly used image processing model. Compared with the traditional artificial neural network model, convolution neural network has more hidden layers. Its unique convolution and pooling operations have higher efficiency in image processing. It has incomparable advantages in image recognition and location and other forms of two-dimensional graphics tasks. As a new deformation of convolution neural network, residual neural network aims to make convolution layer learn a kind of residual instead of a direct learning goal. After analyzing the characteristics of CNN model for image feature representation and residual network, a residual network model is built. The UAV remote sensing system is selected as the platform to acquire image data, and the problem of image recognition based on residual neural network is studied, which is verified by experiment simulation and precision analysis. Finally, the problems and experiences in the process of learning and designing are discussed, and the future improvements in the field of image target location and recognition are prospected.
关键词： Residual network,CNN,Image recognition,UAV
更新于2025-09-23 15:22:29
-
Passive Optical Network Based Mobile Backhaul Enabling Ultra-Low Latency for Communications Among Base Stations
摘要： Low latency is of key importance for mobile networks to support emerging time-critical applications, such as road traffic safety and efficiency. Meanwhile, a passive optical network (PON) is widely recognized as a promising solution for mobile backhaul networks thanks to its high capacity and low energy consumption. In the conventional PON-based mobile backhaul network, where base stations (BSs) are co-located with optical network units, the traffic between the neighboring BSs that are mainly caused by user mobility has to be first sent to the optical line terminal and even further, e.g., edge nodes of mobile core networks, resulting in high latency, although the adjacent BSs are geographically located close to each other. In this paper, a novel PON-based architecture is proposed for mobile backhaul to enhance the connectivity between neighboring BSs. Meanwhile, a tailored medium access control protocol and dynamic bandwidth allocation algorithm are introduced to support fast inter-BS communications. The results reveal that a low latency (less than 1 ms packet delay) for communications among any adjacent BSs can be achieved in the proposed PON-based mobile backhaul network, demonstrating great potential to support future time-critical applications.
关键词： Low latency,Mobile backhaul,Passive optical network,User mobility,Inter-base station communications
更新于2025-09-23 15:22:29
-
3D surface reconstruction of retinal vascular structures
摘要： We propose in this paper, a three-dimensional surface reconstruction of a retinal vascular network from a pair of 2D retinal images. Our approach attempts to address the above challenges by incorporating an epipolar geometry estimation and adaptive surface modelling in a 3D reconstruction, using three steps: segmentation, 3D skeleton reconstruction and 3D surface modelling of vascular structures. The intrinsic calibration matrices are found via the solution of simplified Kruppa equations. A simple essential matrix based on a self-calibration method has been used for the ‘fundus camera-eye’ system. The used method has eventually produced vessel surfaces that could be fit for various applications, such as applications for computational fluid dynamics simulations and applications for real-time virtual interventional.
关键词： Kruppa equations,curvature-dependent subdivision,surface reconstruction,epipolar geometry,segmentation,retinal vascular network,self-calibration
更新于2025-09-23 15:22:29
-
Estimation of Uncertainty in the Lateral Vibration Attenuation of a Beam with Piezo-Elastic Supports by Neural Networks
摘要： Quantification of uncertainty in technical systems is often based on surrogate models of corresponding simulation models. Usually, the underlying simulation model does not describe the reality perfectly, and consequently the surrogate model will be imperfect. In this article we propose an improved surrogate model of the vibration attenuation of a beam with shunted piezoelectric transducers. Therefore, experimentally observed and simulated variations in the vibration attenuation are combined in the model estimation process, by using multi–layer feedforward neural networks. Based on this improved surrogate model, we construct a density estimate of the maximal amplitude in the vibration attenuation. The density estimate is used to analyze the uncertainty in the vibration attenuation, resulting from manufacturing variations.
关键词： Density estimation,neural network,uncertainty quantification,imperfect model,surrogate model
更新于2025-09-23 15:22:29
-
Research on Image Restoration Algorithms Based on BP Neural Network
摘要： With the development of information transmission technology and computer technology, information acquisition mode is mainly converted from character to image nowadays. However, in the process of acquiring and transmitting images, image damage and quality decrease due to various factors. Therefore, how to restore image has become a research hotspot in the field of image processing. This paper establishes an image restoration model based on BP neural network. The simulation results show that the proposed method has made a great improvement compared with the traditional image restoration method.
关键词： image processing,BP neural network,image restoration,image denoising
更新于2025-09-23 15:22:29
-
Energy Harvesting Wireless Communications || Energy Harvesting in Next-Generation Cellular Networks
摘要： To handle the explosive growth of mobile traffic, next-generation cellular network will deploy more and more small-cell BSs (SBSs) in addition to the macro base stations (MBSs). The resultant network, namely, the heterogeneous network (HetNet), provides capacity boost on one hand but brings more energy consumption with the densely deployed SBSs on the other hand. In fact, due to the dynamics of wireless traffic load, many BSs are lightly loaded but almost work at their peak power, due to the elements like power amplifiers and supporting circuits. Unfortunately, these BSs can hardly be turned off for the coverage guarantee. To solve this problem, a new separation architecture called hyper-cellular network (HCN) is proposed, and the main idea is to decouple the function of control signaling from the function of data transmission, such that the data coverage can match the traffic dynamics in a more elastic way. Under HCN, SBSs are only utilized for high data rate transmission, whereas MBSs guarantee the network coverage and provide low data rate service. Therefore, SBSs can be turned off to save energy without worrying about the user coverage. To this end, its nature is to further power SBSs with renewable energy to save more grid power consumption. However, due to the randomness of renewable energy arrivals, it is challenging to manage wireless resource and the on-off states of energy harvesting (EH) BSs. It can be more challenging in HCN. First, diverse types of SBSs may be equipped with different kinds of energy sources, making the energy arrival statistically nonuniform over the space. In addition, the traffic load is nonevenly distributed across different base station (BS) tiers and also not in accordance with the energy arrivals over the spatial and temporal domains. To this end, on top of the techniques introduced in Chapter 4, in HCN the key to match the random energy arrival with the traffic load variation over time and space is to jointly optimize the working states of SBSs and the user traffic offloading. Although traffic offloading has been extensively studied in grid-powered cellular networks, the conventional offloading methods cannot be directly applied as they do not consider the energy states of BSs. Accordingly, energy-aware traffic offloading schemes are needed, and some energy-aware traffic offloading schemes have been proposed for single-tier homogeneous networks and two-tier HCN with one renewable energy-powered SBS, respectively. In the first part of the chapter, we will illustrate how to coordinate the on-off switching of SBSs with inter-tier traffic offloading, under the scenario with different types of SBSs, powered by various energy sources. The goal is to minimize the on-grid power consumption of the whole HCN system while satisfying the quality of service (QoS) requirements of users. Another emerging technology of next-generation cellular networks is to exploit edge caching with proactive services, like push. While the initial motivation of proactive caching and push is to reduce the duplicated content transmissions, and thus reduce the core network traffic load as well as the content delivery delay, it is also beneficial to address the mismatch between the energy and traffic in renewable energy-powered SBSs. Specifically, the contents can be cached at the storage of SBSs and then pushed to users earlier than the actual demands when there is sufficient harvested energy. The users can successfully get the contents when they actually require it even if at that time the SBS does not have enough energy for transmission. Consequently, the energy waste due to the battery overflow can be avoided as the harvested energy can be used effectively and timely. It can be viewed as transferring the harvested energy along with the timeline to the future to match the random energy arrival with the traffic needs. In the second part of this chapter, we will demonstrate the concept of integrating proactive service provisioning with EH HCN and provide a detailed study on the optimal policy design for content push from an EH-based SBS.
关键词： push,cellular networks,renewable energy,proactive caching,traffic offloading,quality of service,small-cell base stations,Markov decision process,Energy harvesting,hyper-cellular network
更新于2025-09-23 15:22:29
-
An Attribute-based High-level Image Representation for Scene Classification
摘要： Scene classification is increasingly popular due to its extensive usage in many real-world applications such as object detection, image retrieval, and so on. Traditionally, the low-level hand-crafted image representations are adopted to describe the scene images. However, they usually fail to detect semantic features of visual concepts, especially in handling complex scenes. In this paper, we propose a novel high-level image representation which utilizes image attributes as features for scene classification. More specifically, the attributes of each image are firstly extracted by a deep convolution neural network (CNN), which is trained to be a multi-label classifier by minimizing an element-wise logistic loss function. The process of generating attributes can reduce the 'semantic gap' between the low-level feature representation and the high level scene meaning. Based on the attributes, we then build a system to discover semantically meaningful descriptions of the scene classes. Extensive experiments on four large-scale scene classification datasets show that our proposed algorithm considerably outperforms other state-of-the-art methods.
关键词： high-level image representation,Scene classification,attribute representation,convolutional neural network
更新于2025-09-23 15:22:29