Application of neural networks for sea ice classification inpolarimetric SAR images

Several automatic methods have been developed to classify sea ice types from fully polarimetric synthetic aperture radar (SAR) images, and these techniques are generally grouped into supervised and unsupervised approaches. In previous work, supervised methods have been shown to yield higher accuracy than unsupervised techniques, but suffer from the need for human interaction to determine classes and training regions. In contrast, unsupervised methods determine classes automatically, but generally show limited ability to accurately divide terrain into natural classes. In this paper, a new classification technique is applied to determine sea ice types in polarimetric and multifrequency SAR images, utilizing an unsupervised neural network to provide automatic classification, and employing an iterative algorithm to improve the performance. The learning vector quantization (LVQ) is first applied to the unsupervised classification of SAR images, and the results are compared with those of a conventional technique, the migrating means method. Results show that LVQ outperforms the migrating means method, but performance is still poor. An iterative algorithm is then applied where the SAR image is reclassified using the maximum likelihood (ML) classifier. It is shown that this algorithm converges, and significantly improves classification accuracy. The new algorithm successfully identifies first-year and multiyear sea ice regions in the images at three frequencies. The results show that L- and P-band images have similar characteristics, while the C-band image is substantially different. Classification based on single features is also carried out using LVQ and the iterative ML method. It is found that the fully polarimetric classification provides a higher accuracy than those based on a single feature. The significance of multilook classification is demonstrated by comparing the results obtained using four-look and single-look classifications     

A Bayesian classification model for sea ice roughness fromscatterometer data

For sea ice in the Baltic Sea, surface scattering can be regarded as the dominant scattering mechanism at C-band. In this paper, a new statistical method is introduced for making statistical inferences about the underlying ice surface roughness on the basis of one-dimensional (1D) scatterometer data y. The central parameter in the hierarchical model applied in the context is a mixture parameter p, which indicates the degree of surface roughness in ice surface. Several questions related to the occurrence of different ice classes on a transect can be solved with the aid of the posterior distribution [p|y]. An empirical approximation for the posterior distribution is computed by using Markov Chain Monte Carlo methodology. The efficiency of the suggested approach is investigated by analyzing a C-band HH-polarization helicopter-borne HUTSCAT scatterometer data. The results provided by the statistical model show good agreement with a video-based ice type classification     

基于前向神经网络的音乐情感识别分类算法

为解决单模态数据在音乐情感分类上的局限性,并同时提高对音乐情感分类的准确性,文中提出了一种基于前向神经网络的多特征融合音乐分类算法。在传统的前向神经网络模型中融入切比雪夫正交多项式簇作为隐藏层各神经元的激励函数,使每一层神经元的激励函数各不相同。利用梯度下降学习算法来进行网络参数的有监督训练;同时利用音频、歌词中不同模态的数据,使其形成多模态数据,来进行音乐情感分类模型的训练。实验测试结果表明,该算法对音乐情感的分类具有较好的效果,平均准确率为78. 37%,具有良好的有效性与可行性。     

基于神经网络的红外目标分类算法设计与应用

卷积神经网络在图像处理领域取得了突出表现,但是由于算法庞大的计算量引起功耗高和实时性差的问题导致神经网络的实际应用受到一定限制。如果将神经网络移植在FPGA硬件平台,可充分发挥其高度并行的优势实现网络加速,降低功耗并提升算法实时性。基于上述描述,本文将用于目标分类的网络模型成功移植在FPGA上,通过对比加入分类模型前后的告警结果,说明分类模型设计的重要性。对比硬件实现与仿真结果,证明硬件实现的正确性。     

基于部分连接神经网络的序列数据分类算法研究

研究梳理了序列数据的定义,并且分析了5种类型的序列数据,结合局部连接神经网络的特点,研究了基于局部连接神经网络的序列数据的分类算法。使用该算法进行计算,学习与收敛速度较快,对于自适应建模与控制十分适用,利用方形基函数进行计算,在网络输出过程中注意只能利用方形函数来逼近光滑函数。由于序列数据分类运算在数据挖掘中存在巨大的优势,因此对序列数据算法的研究具有很高的理论与应用价值。     

遥感图像分类中的遗传算法LVQ神经网络运用

对于遥感图像分类过程中的问题,提出遗传算法LVQ神经网络来实现遥感图像的分类。将LVQ神经网络结合遗传算法,使用遗传算法最优阈值与权值实现网络训练,使分类精度得到提高。之后融合相似灰度值创建分类图像特征矢量,使特征矢量在神经网络中输入实现训练。学习矢量量化神经算法对初值非常敏感,对遥感图像分类精度具有一定影响。最后,为了对性能进行测试,在实验过程中对比本文分类方法和SVM决策树分类方法,通过实验结果表示,文中提出的分类方法的遥感图像分类精度为95.82%,与其他分类方法相比,分类精度得到进一步提高。     

Fuzzy-ChOA: an improved chimp optimization algorithm for marine mammal classification using artificial neural network

Analog Integrated Circuits and Signal Processing - Chimp optimization algorithm (ChOA) is a robust nature-inspired technique, which was recently proposed for addressing real-world challenging...     

TBNN: totally-binary neural network for image classification

Most binary networks apply full precision convolution at the first layer. Changing the first layer to the binary convolution will result in a significant loss of accuracy. In this paper, we propose a new approach to solve this problem by widening the data channel to reduce the information loss of the first convolutional input through the sign function. In addition, widening the channel increases the computation of the first convolution layer, and the problem is solved by using group convolution. The experimental results show that the accuracy of applying this paper''s method to state-of-the-art (SOTA) binarization method is significantly improved, proving that this paper''s method is effective and feasible.     

面向室内定位的DHOHF-Elman神经网络算法

针对传统Elman神经网络算法在室内存在定位精度低的问题,提出了一种基于UWB（Ultra Wideband ）的改进的DHOHF-Elman(Elman neural network with Double Hidden layers and Output-Hidden Feedback,DHOHF-Elman)神经网络算法。该算法改进了神经网络拓扑结构增加了第二隐含层和第二承接层,达到了双隐含层反馈的效果,采集了大量的实验数据对构造的神经网络模型进行了训练与测试,表明了改进后的神经网络算法较传统神经网络算法有更高的定位精度和较好的收敛性,最后通过仿真结果分析验证了改进算法的优良性和有效性。      

基于语义特征传播图神经网络的小样本图像分类算法

使用少量样本进行学习和概括的能力是人工智能和人类之间主要的区别。在小样本学习领域,大多数图神经网络专注于将标记的样本信息传递给未标记的查询样本,而忽略了语义特征在分类过程中的重要作用。为此构建了语义特征传播图神经网络,首先将语义特征嵌入到图神经网络中,解决了细粒度图像特征相似性带来的分类准确率低的问题,然后将注意力机制与骨干网络合并达到强化前景并提高特征提取质量的目的,利用马氏距离计算类的相似度得到更好的分类性能,最后使用Funnel ReLU函数作为激活函数进一步提高分类准确率。在基准数据集上实验表明,所提算法相比于基线算法在5类1/2/5样本任务上的准确率分别提高了903%、456%和415%。     

A scattering model for snow-covered sea ice

The special properties of a robust radiative transfer model for scattering from layers of inhomogeneous rough-boundary slabs are presented. The model is applied to backscattering from saline and desalinated ice. Comparisons are made at single and multiple frequencies with some of the most complete sets of measurement data available, using measured physical and electrical characteristics of the ice as inputs to the model where possible. The results show close agreement. For example, for the saline ice backscatter data set, which consisted of measurements at two like and two cross polarizations at 5 and 13.9 GHz, the agreement with model predictions is within 2 dB except at 13.9-GHz cross polarization. Backscattering from >15-cm-thick saline ice is generally dominated by scattering from the top surface while backscattering from <8-cm-thick saline ice can be strongly influenced by returns from the ice/water interface, particularly at frequencies less than about 5 GHz     

A hybrid fuzzy logic and neural network algorithm for robot motioncontrol

Robotic manipulators are multivariable nonlinear coupling dynamic systems. Industrial robots were controlled by using a traditional controller, the control performance of which may change with respect to operating conditions. Since the robotic manipulators have complicated nonlinear mathematical models, control systems based on the system model are difficult to design. In this paper, a model-free hybrid fuzzy logic and neural network algorithm was proposed to control this multi-input/multi-output (MIMO) robotic system. First, a fuzzy logic controller was designed to control individual joints of this 4-degree-of-freedom (DOF) robot. Secondly, a coupling neural network controller was introduced to take care of the coupling effect among joints and refine the control performance of this robotic system. The experimental results showed that the application of this control strategy effectively improved the trajectory tracking precision     

Randomized neural network based signature for color texture classification

Multidimensional Systems and Signal Processing - Color texture analysis is an important subject in computer vision research. This paper presents an innovative and powerful color texture analysis...     

Multisensor approach to automated classification of sea ice image data

A multisensor data fusion algorithm based on a multilayer neural network is presented for sea ice classification in the winter period. The algorithm uses European Remote Sensing (ERS), RADARSAT synthetic aperture radar (SAR), and low-resolution television camera images and image texture features. Based on a set of in situ observations made at the Kara Sea, a neural network is trained, and its structure is optimized using a pruning method. The performance of the algorithm with different combinations of input features (sensors) is assessed and compared with the performance of a linear discriminant analysis (LDA)-based algorithm. We show that for both algorithms a substantial improvement can be gained by fusion of the three different types of data (91.2% for the neural network) as compared with single-source ERS (66.0%) and RADARSAT (70.7%) SAR image classification. Incorporation of texture increases classification accuracy. This positive effect of texture becomes weaker with increasing number of sensors (from 8.4 to 6.4 percent points for the use of two and three sensors, respectively). In view of the short training time and smaller number of adjustable parameters, this result suggests that semiparametric classification methods can be considered as a good alternative to the neural networks and traditional parametric statistical classifiers applied for the sea ice classification.     

基于神经网络的海面溢油图像分类算法

提出了一种基于Gabor变换、KPCA和神经网络的图像分类方法。首先对图像进行Gabor滤波,获得不同方向的特征参数;然后提取图像的KPCA作为图像的特征,最后利用神经网络进行分类。通过对实验分类结果的定量分析可知,该方法可以获得精度比最小分类模型方法以及最大似然分布模型方法高的分类结果。     

RBF神经网络LM算法的改进

文章对现有的RBF神经网络算法进行改进,改进的基本思想是：采用L-M算法训练RBF网络,并对L-M算法的重要参数提出一种随迭代步数的动态调整方法,从而提高运算的精度和效率,并经过仿真验证了提出改进方案的有效性。     

基于门控卷积神经网络的图像超分辨重建算法北大核心CSCD

近年来,卷积神经网络被广泛应用于图像超分辨率领域。针对基于卷积神经网络的超分辨率算法存在图像特征提取不充分,参数量大和训练难度大等问题,本文提出了一种基于门控卷积神经网络(gated convolutional neural network, GCNN)的轻量级图像超分辨率重建算法。首先,通过卷积操作对原始低分辨率图像进行浅层特征提取。之后,通过门控残差块(gated residual block, GRB)和长短残差连接充分提取图像特征,其高效的结构也能加速网络训练过程。GRB中的门控单元(gated unit, GU)使用区域自注意力机制提取输入特征图中的每个特征点权值,紧接着将门控权值与输入特征逐元素相乘作为GU输出。最后,使用亚像素卷积和卷积模块重建出高分辨率图像。在Set14、BSD100、Urban100和Manga109数据集上进行实验,并和经典方法进行对比,本文算法有更高的峰值信噪比(peak signal-to-noise ratio,PSNR)和结构相似性(structural similarity,SSIM),重建出的图像有更清晰的轮廓边缘和细节信息。