基于深度卷积神经网络的协作频谱感知方法

针对传统卷积神经网络(CNN)频谱感知方法提取特征能力受限于网络结构简单,增加网络结构又容易出现梯度消失等问题,该文通过在传统卷积神经网络中添加捷径连接,实现输入层恒等映射更深的网络,提出一种基于深度卷积神经网络(DCNN)的协作频谱感知方法。该方法将频谱感知问题转化为图像二分类问题,对正交相移键控(QPSK)信号的协方差矩阵进行归一化灰度处理,并作为深度卷积神经网络的输入,通过残差学习训练深度卷积神经网络模型,提取2维灰度图像的深层特征,将测试数据输入到训练好的模型中,完成基于图像分类的频谱感知。实验结果表明:与传统的频谱感知方法相比,在低信噪比(SNR)下、多用户协作感知时,所提方法具有更高的检测概率和更低的虚警概率。     

Evolutionary Extreme Learning Machine and Its Application to Image Analysis

Extreme learning machine (ELM) and evolutionary ELM (E-ELM) were proposed as a new class of learning algorithm for single-hidden layer feedforward neural network (SLFN). In order to achieve good generalization performance, E-ELM calculates the error on a subset of testing data for parameter optimization. Since E-ELMemploys extra data for validation to avoid the overfitting problem, more samples are needed for model training. In this paper, the cross-validation strategy is proposed to be embedded into the training phase so as to solve the overtraining problem. Based on this new learning structure, two extensions of E-ELM are introduced. Experimental results demonstrate that the proposed algorithms are efficient for image analysis.     

基于残差网络的光学遥感图像场景分类算法

提出一种基于卷积神经网络中残差网络的遥感图像场景分类方法。本文方法在原网络模型中嵌入了跳跃连接和协方差池化两个模块,用于连接多分辨率特征映射和融合不同层次的多分辨率特征信息,并在3个公开的经典遥感数据集上进行了实验。结果证明,本文方法不仅可以将残差网络中不同层次的多分辨率特征信息融合在一起,还可以利用高阶信息来实现更具代表性的特征学习。与已有的分类方法相比,本文方法在场景分类问题上拥有更高的分类精度。     

A pruning ensemble model of extreme learning machine with $$L_{1/2}$$ regularizer

Extreme learning machine (ELM) as an emerging branch of machine learning has shown its good generalization performance at a very fast learning speed. Nevertheless, the preliminary ELM and other evolutional versions based on ELM cannot provide the optimal solution of parameters between the hidden and output layer and cannot determine the suitable number of hidden nodes automatically. In this paper, a pruning ensemble model of ELM with \(L_{1/2} \) regularizer (PE-ELMR) is proposed to solve above problems. It involves two stages. First, we replace the original solving method of the output parameter in ELM to a minimum squared-error problem with sparse solution by combining ELM with \(L_{1/2}\) regularizer. Second, in order to get the required minimum number for good performance, we prune the nodes in hidden layer with the ensemble model, which reflects the superiority in searching the reasonable hidden nodes. Experimental results present the good performance of our method PE-ELMR, compared with ELM, OP-ELM and PE-ELMR (L1), for regression and classification problems under a variety of benchmark datasets.     

基于深度分层特征表示的行人识别方法

该文针对行人识别中的特征表示问题,提出一种混合结构的分层特征表示方法,这种混合结构结合了具有表示能力的词袋结构和学习适应性的深度分层结构。首先利用基于梯度的HOG局部描述符提取局部特征,再通过一个由空间聚集受限玻尔兹曼机组成的深度分层编码方法进行编码。对于每个编码层,利用稀疏性和选择性正则化进行无监督受限玻尔兹曼机学习,再应用监督微调来增强分类任务中视觉特征表示,采用最大池化和空间金字塔方法得到高层图像特征表示。最后采用线性支持向量机进行行人识别,提取深度分层特征遮挡等与目标无关部分自然分离,有效提高了后续识别的准确性。实验结果证明了所提出方法具有较高的识别率。     

Joint Feature Representation and Classifier Learning Based Unsupervised Domain Adaption ELM

In the problem of unsupervised domain adaption Extreme learning machine (ELM), the output layer parameters need to have both classification and domain adaptation functions, which often cannot be simultaneously fully utilized. In addition, traditional matching method based on data probability distribution cannot find the common subspace of source and target domains under large difference between domains. In order to alleviate the pressure of double functions of classifier parameters, the entire ELM learning process is mainly divided into two stages: feature representation and adaptive classifier learning, thus a joint feature representation and classifier learning based unsupervised domain adaption ELM model is proposed. In the feature representation stage, the source and target domain data are projected to their respective subspace while minimizing the difference in probability distribution between the two domains. In the adaptive classifier learning stage, the smooth manifold regularization term of target domain is used to improve the parameter adaptive ability. Experiments on six different types of datasets show that the proposed model has higher cross-domain classification accuracy.     

基于HOGv-CLBP特征融合和ELM的交通标志识别

针对交通标志识别识别率低和时间复杂度大的问题 ,本文提出一种HOGv-CLBP特征融合和极限学习机的交通标志识别算法。首先通过描述交通 标志图像边缘信息的方向梯度直方图(HOG)特征,与能够 表示标志图像内部纹理信息的局部二值模式(LBP)特征融合得到降维后形成一种HOGv-CLB P有效特征, 然后利用ELM进行交通标志训练和分类。实验结果表明,该算法不仅提高了交通标志的识别 率,而且降低了时间复杂度,增强了系统鲁棒性。     

联合多尺度多特征的高分遥感图像场景分类

高分辨率遥感图像地物信息丰富,但场景构成复杂,目前基于手工设计的特征提取方法不能满足复杂场景分类的需求,而非监督特征学习方法尽管能够挖掘局部图像块的本征结构,但单一种类及尺度的特征难以有效表达实际应用中复杂遥感场景特性,导致分类性能受限.针对此问题,本文提出了一种基于多尺度多特征的遥感场景分类方法.该算法首先设计了一种改进的谱聚类非监督特征（iUFL-SC）以有效表征图像块的本征结构,然后通过密集采样提取每幅遥感场景的iUFL-SC、LBP、SIFT等三种多尺度局部图像块特征,并通过视觉词袋模型（BoVW）获得场景的中层特征表达,以实现更为准确详实的特征描述,最后基于直方图交叉核的支持向量机（HIKSVM）进行分类.在UC Merced数据集以及WHU-RS19数据集上的实验结果表明本文方法可对遥感场景进行鉴别特征提取,有效提高分类性能.     

结合多特征融合和极限学习机的红外图像目标分类方法

针对红外图像目标分类问题,提出了结合多特征融合和极限学习机（extreme learning machine,ELM）的方法。采用主成分分析（principal component analysis,PCA）、局部二值模式（local binary pattern,LBP）以及尺度不变特征变换（scale-invariant feature transform,SIFT）三类特征分别描述红外图像中目标的像素分布、局部纹理以及特征点信息。三类特征从不同侧面反映红外图像目标特性,因此具有互为补充的优势。在此基础上,基于多重集典型相关分析（multiset canonical correlations analysis,MCCA）对三类特征进行融合处理,获得统一的特征矢量。融合后的特征不仅继承了原始三类特征的鉴别特性,还有效去除了冗余信息。分类过程中,采用极限学习机作为基础分类器对融合特征矢量进行分类。极限学习机具有参数少、效率高、精度高和稳健性强等显著特点,有利于提高红外目标分类的整体性能。因此,所提出的方法通过结合多特征和极限学习机的优势综合提升了目标识别性能。在实验过程中,采用四类飞机目标的红外图像对所提出方法进行了性能测试。根据与现有几类方法的对比,实验结果证明了提出方法的性能优势。     

基于深度卷积网络的高分遥感图像语义分割

随着我国高分对地观测系统的不断发展,对高分影像智能化分析与处理的应用需求愈来愈多,基于深度学习语义分割的影像分类也受到高度关注。作为近景图像语义分割的热点模型,Deeplab网络在应用时取得了良好的效果。为了解决多尺度高分辨率遥感影像语义分割问题,本文首先利用空洞卷积扩大Atrous空间金字塔池化（ASPP）结构的感受野,然后对DeepLabv3进行改进并将其用于高分2号遥感影像的分类处理。我们以郴州地区的高分遥感影像为研究对方法进行了验证,首先对原始影像进行预处理,再对预处理图像进行数据增强与扩充,最后通过对不同参数条件下的分类结果进行对比,分析该模型的适应性和精确性。在我们的数据集中,本文方法的实验分类像素精度为88.2％,MIoU达到72.5％,得到了比Deeplab更好的分类效果。      

基于空间约束多特征联合稀疏编码的遥感图像标注方法研究

针对稀疏表示分类器不能较好地适应多特征框架的问题,该文提出一种空间约束多特征联合稀疏编码模型,并以此实现遥感影像的自动标注。该方法利用l1,2混合范数正则化多特征编码系数,约束编码系数共享相同的稀疏模式,在保持多特征关联的同时,又不添加过于严格的约束。同时,将字典学习技术扩展到多特征框架中,通过约束字典更新的变换矩阵,解决了字典学习过程丢失多特征关联的问题。另外,针对遥感影像中的空间关系常常被忽略或者利用不充分的不足,还提出了将空间一致性与多特征联合稀疏编码相结合的分类准则,提高了标注性能。在遥感公开数据集与大尺寸卫星影像上的实验证明了该方法的有效性。     

改进贝叶斯ARTMAP的迁移学习遥感影像分类算法

遥感分类旨在从图像光谱中提取资源环境监测可用的地理信息,然而基于模式分类的图像处理技术受光谱漂移影响而缺乏历史样本重复利用的有效策略,制约着有限目标样本下遥感分类精度的提高.针对该问题,本文构建了基于改进的贝叶斯ARTMAP神经网络的迁移学习遥感影像分类算法,通过提高谐振匹配性来抑制类别扩散,利用节点的离散增量期望最大化参数更新策略,将历史遥感样本中的地物分类先验信息迁移到目标模型当中.实验结果表明本文方法能有效利用历史遥感数据弥补缺少目标训练数据的不足,相比于其他样本利用策略大幅提高遥感影像分类精度.     

基于深度学习的红外遥感信息自动提取

为了提高红外遥感图像地物 信息自动提取的精确性,同时避免人工提取遥感 信息的低效性,提出了一种基于UNet深度学习模型 的遥感信息提取算法。该算法用于从红外遥感图像中分割 出5类地物信息(包括道路、建筑、树木、农田和水 体)。首先,对分辨率高但数量较少的训练数 据进行小像幅的随机裁剪,并对其进行相应的数据增 强处理。然后搭建UNet深度学习模型,并用它 自动提取遥感图像的特征信息。采用交叉熵损失函数 以及Adam反向传播优化算法对该模型进行训练,并对测 试样本中的5幅遥感图像进行精确的地物信息提取。最后,运 用Jaccard指数对测试结果进行精度评定。实验结果表明,该 方法对高分辨率红外遥感图像信息和可见光 遥感图像信息进行了充分融合,对于不同种类地物 的定位和分类都取得了较高精度。     

Regularized manifold information extreme learning machine

By exploiting the thought of manifold learning and its theoretical method, a regularized manifold information ex-treme learning machine algorithm aimed to depict and fully utilize manifold information was proposed. The proposed algo-rithm exploited the geometry and discrimination manifold information of data to perform network of ELM. The proposed algorithm could overcome the problem of the overlap of information. Singular problems of inter-class and within-class were solved effectively by using maximum margin criterion. The problem of inadequate learning with limited samples was solved. In order to demonstrate the effectiveness, comparative experiments with ELM and the related update algorithms RAFELM, GELM were conducted using the commonly used image data. Experimental results show that the proposed algorithm can significantly improve the generalization performance of ELM and outperforms the related update algorithms.     

一种稳定的基于极端学习机的纹理分类方法

针对传统基于极端学习机(ELM)的纹理分类方 法容易出现输出不稳定的缺点,本文将线性和非线 性模型进行融合,改进了传统动力学模型。利用ELM能够快速学习的特点,将其作为基分类 器,同时利用 线性吸引子和局部吸引子的迭代,实现多个ELM分类器的最佳融合,以提升学习模型的泛 化能力和稳定 性。利用动力模型实现多分类器的融合有助于寻求多个基分类器之间的一致性,摒弃了基分 类器中判别错 误的样本输出。通过对CUReT纹理数据库的分类结果,与传统纹理分类方法相 比,本文算法的稳定性 和分类准确率都得到了一定程度的提升,取得了理想的分类效果。     

Collaborative Spectrum Sensing for Illegal Drone Detection: A Deep Learning-Based Image Classification Perspective

Drones,also known as mini-unmanned aerial vehicles(UAVs),are enjoying great popularity in recent years due to their advantages of low cost,easy to pilot and small size,which also makes them hard to detect.They can provide real time situational awareness information by live videos or high definition pictures and pose serious threats to public security.In this article,we combine collaborative spectrum sensing with deep learning to effectively detect potential illegal drones with states of high uncertainty.First,we formulate the detection of potential illegal drones under illegitimate access and rogue power emission as a quaternary hypothesis test problem.Then,we propose an algorithm of image classification based on convolutional neural network which converts the cooperative spectrum sensing data at a sensing slot into one image.Furthermore,to exploit more information and improve the detection performance,we develop a trajectory classification algorithm which converts theflight process of the drones in consecutive multiple sensing slots into trajectory images.In addition,simulations are provided to verify the proposed methods’performance under various parameter configurations.     

An adaptive method for combined covariance estimation andclassification

In this paper, a family of adaptive covariance estimators is proposed to mitigate the problem of limited training samples for application to hyperspectral data analysis in quadratic maximum likelihood classification. These estimators are the combination of adaptive classification procedures and regularized covariance estimators. In these proposed estimators, the semi-labeled samples (whose labels are determined by a decision rule) are incorporated in the process of determining the optimal regularized parameters and estimating those supportive covariance matrices that formulate final regularized covariance estimators. In all experiments with simulated and real remote sensing data, these proposed combined covariance estimators achieved significant improvement on statistics estimation and classification accuracy over conventional regularized covariance estimators and an adaptive maximum likelihood classifier (MLC). The degree of improvement increases with dimensions, especially for ill-posed or very ill-posed problems where the total number of training samples is smaller than the number of dimensions