- 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
- Optoelectronic Information Science and Engineering
- Wuhan University
- Central South University
- Hubei University
-
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
-
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
-
[IEEE 2018 IEEE Power & Energy Society General Meeting (PESGM) - Portland, OR, USA (2018.8.5-2018.8.10)] 2018 IEEE Power & Energy Society General Meeting (PESGM) - Study of Impact of Cloud Distribution on Multiple Interconnected Solar PV Plants Generation and System Strength
摘要： Dependence of solar power generation on solar irradiance results in sudden and dramatic variations in power generation following significant changes in cloud distribution over a solar PV plant. Currently, this phenomenon is being one of the most challenging issues in resource planning and maintaining the reliability of modern power grids with high penetration of solar power. The dramatic variation of solar power generation has a direct impact on system strength at the Points of Interconnection (POIs). Hence, the power quality of the system is compromised, especially because solar PV plants are usually interconnected to distribution systems and near load zones. In this paper, an Artificial Neural Network (ANN) based approach is developed to forecast the clouds distribution for the estimation of sudden and dramatic variations in the solar irradiance. This estimate is used to evaluate the system strength in terms of voltage stability at each POI. We apply newly developed methodology to measure the system strength known as Site-Dependent Short Circuit Ratio (SDSCR), which provides more accurate results of system strength evaluation. The validity and effectiveness of the developed approach is confirmed through comparing its results versus the cloud distribution data provided by weather satellites.
关键词： Artificial neural network,renewable energy,system strength,voltage stability,short circuit ratio
更新于2025-09-23 15:22:29
-
Metric Learning for Patch-Based 3-D Image Registration
摘要： Patch-based image registration is a challenging problem in visual geometry, the crucial component of which is the selection of an appropriate similarity measure. The similarity measure participates in the objective calculation of the pose optimization, which determines the optimization convergence performance. In this paper, we propose learning a similarity metric of patches from reference and target images such that the pairwise patches with a small projection error receive high similarity scores. To achieve this objective, we designed and trained the classification, regression, and rank networks separately based on self-collected data sets. The network can directly output the projection error according to the patches, which is sensitive to the deviation of the pose transformation. We also designed evaluation criteria and validated the superior performance of the network's outputs compared with the performance of traditional methods, such as the sum of absolute difference and the sum of squared differences.
关键词： neural network,Image registration,pose optimization,metric learning
更新于2025-09-23 15:22:29
-
Sky Image-Based Solar Irradiance Prediction Methodologies Using Artificial Neural Networks
摘要： In order to decelerate global warming, it is important to promote renewable energy technologies. Solar energy, which is one of the most promising renewable energy sources, can be converted into electricity by using photovoltaic power generation systems. Whether the photovoltaic power generation systems are connected to an electrical grid or not, predicting near-future global solar radiation is useful to balance electricity supply and demand. In this work, two methodologies utilizing artificial neural networks (ANNs) to predict global horizontal irradiance in 1 to 5 minutes in advance from sky images are proposed. These methodologies do not require cloud detection techniques. Sky photo image data have been used to detect the clouds in the existing techniques, while color information at limited number of sampling points in the images are used in the proposed methodologies. The proposed methodologies are able to capture the trends of fluctuating solar irradiance with minor discrepancies. The minimum root mean square errors of 143 W/m2, which are comparable with the existing prediction techniques, are achieved for both of the methodologies. At the same time, the proposed methodologies require much less image data to be handled compared to the existing techniques.
关键词： Artificial Neural Network,Photovoltaic Power Generation,Solar Energy,Global Horizontal Irradiance Prediction,Sky Image
更新于2025-09-23 15:21:21
-
A novel method on the edge detection of infrared image
摘要： Infrared image processing is important for fault identification of high-voltage equipment. This paper studies the problem on the edge detection of infrared image. First a kind of spiking neural network is constructed, and by using the characteristics of the spiking neuron, a novel method is designed to achieve the edge detection of infrared image. Finally, some typical examples are included and corresponding experimental results show the effectiveness and advantage of the proposed method.
关键词： spiking neural network,Edge detection,infrared image
更新于2025-09-23 15:21:21
-
Star sensor installation error calibration in stellar-inertial navigation system with a regularized backpropagation neural network
摘要： The star sensor is the attitude reference in a stellar-inertial navigation system. It is essential to acquire the star sensor installation error, which has a great influence on the system navigation performance. However, traditional methods have a poor tolerance for a large range of installation errors, especially when the system works under a separate installation mode. In this paper a novel calibration method, using a regularized backpropagation (BP) neural network, is proposed. With a specially designed calibration procedure, the neural network is structured with BP and the regularization is improved. The network training is conducted for parameter solidification. The calibration can be achieved without formula derivation and numerical calculation under both small and large installation errors. In the experiment, the calibration accuracy is about 5 arcsec under small installation errors and about 20 arcsec under large installation errors, which is much better than a Kalman filter. The proposed method has the potential to be a universal star sensor calibration method under integrative installation mode or separated installation mode with large installation error.
关键词： neural network,installation error calibration,star sensor
更新于2025-09-23 15:21:21
-
[IEEE 2018 15th European Radar Conference (EuRAD) - Madrid, Spain (2018.9.26-2018.9.28)] 2018 15th European Radar Conference (EuRAD) - Deep Learning based Human Activity Classification in Radar Micro-Doppler Image
摘要： A convolutional neural network (CNN) based deep learning (DL) approach to classify human activities in micro-Doppler spectrogram of radar is investigated. MOCAP dataset, from Carnegie Mellon University, is used for spectrogram simulation. Seven activities are classified with the proposed CNN network. Our network outperforms several previously published DL-based approaches. To understand the network’s impact on classification performance, we investigate some key parameters of the proposed network. Experiment result demonstrates that a deeper network does not necessarily result in a higher accuracy. We also examine the network size and the number of output feature maps to find out their impact on the result.
关键词： Deep Learning,Convolutional Neural Network,Human Activity Classification,Micro-Doppler Spectrogram,Radar image
更新于2025-09-23 15:21:21
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Multitask Classification of Remote Sensing Scenes Using Deep Neural Networks
摘要： The problem of scene classification in remote sensing (RS) images has attracted a lot of attention recently. Many datasets have been presented in the literature for this purpose with each claiming to be the benchmark dataset. In this paper, we propose a different approach to the RS community. Instead of putting our effort in building larger and large scene datasets, we argue that it is better to build a machine learning framework that can learn from all available datasets. We formulate this as a multitask learning problem where each dataset represents a task. Then, we present a deep learning solution that can perform multitask learning. We test the proposed multitask network on three popular scene datasets, namely UC Merced, KSA, and AID datasets. Preliminary results show the promising capabilities of this solution at sharing information between tasks and improving the classification accuracy.
关键词： Deep learning,Scene classification,Multitask classification,Convolutional Neural Network
更新于2025-09-23 15:21:21
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Time-Scale Transferring Deep Convolutional Neural Network for Mapping Early Rice
摘要： In recent years, the use of deep learning in remote sensing domain has made it possible to automate mapping in large-scale. In this paper, we propose a transfer learning method which pre-train a convolutional neural network (CNN) with middle-resolution remote sensing data in 2016, and fine-tune it in following years with a spot of high-resolution remote sensing data in 2017. We used the fine-tuned model to mapping the early-rice in 25 countries which cost only 21 minutes, and yielded an overall accuracy of 81.68%. The result demonstrate that the convolutional neural network model can transfer in different time period with little adjustment in a very high accuracy.
关键词： middle-resolution data,convolutional neural network,time-scale,transfer learning
更新于2025-09-23 15:21:21