Application of SVM on satellite images to detect hotspots in Jharia coal field region of India

The present paper deals with the application of Support Vector Machine (SVM) and image analysis techniques on NOAA/AVHRR satellite image to detect hotspots on the Jharia coal field region of India. One of the major advantages of using these satellite data is that the data are free with very good temporal resolution; while, one drawback is that these have low spatial resolution (i.e., approximately 1.1 km at nadir). Therefore, it is important to do research by applying some efficient optimization techniques along with the image analysis techniques to rectify these drawbacks and use satellite images for efficient hotspot detection and monitoring. For this purpose, SVM and multi-threshold techniques are explored for hotspot detection. The multi-threshold algorithm is developed to remove the cloud coverage from the land coverage. This algorithm also highlights the hotspots or fire spots in the suspected regions. SVM has the advantage over multi-thresholding technique that it can learn patterns from the examples and therefore is used to optimize the performance by removing the false points which are highlighted in the threshold technique. Both approaches can be used separately or in combination depending on the size of the image. The RBF (Radial Basis Function) kernel is used in training of three sets of inputs: brightness temperature of channel 3, Normalized Difference Vegetation Index (NDVI) and Global Environment Monitoring Index (GEMI), respectively. This makes a classified image in the output that highlights the hotspot and non-hotspot pixels. The performance of the SVM is also compared with the performance obtained from the neural networks and SVM appears to detect hotspots more accurately (greater than 91% classification accuracy) with lesser false alarm rate. The results obtained are found to be in good agreement with the ground based observations of the hotspots. This type of work will be quite helpful in the near future to develop a hotspots monitoring system using these operational satellites data.     

一种基于支持向量机的小波图像压缩方法

提出了一种结合支持向量机(SVM,Support Vector Machines)回归与小波变换的新的静态图像压缩方法.SVM回归方法可以学习原始数据之间的相关性,并采用小部分训练样本,即支持向量来稀疏表示原始数据集,利用这一特性来逼近和约减小波系数,可以达到数据压缩的效果.首先采用小波变换把原始图像分解成不同尺度的多个子带,由于最低频子带系数非常重要,采用DPCM直接编码,然后对其它频带系数采用SVM回归进行压缩.由于不同尺度和方向的小波系数特征不同,为尽可能去除小波系数间的各种相关性,给出了适合SVM回归的小波系数的有效组织方式.最后研究了支持向量及其相应权重的混合编码方法.实验结果表明:与同类压缩方法相比,本算法获得的恢复图像的主客观质量有明显提高.      

基于支持向量机的机场智能驱鸟决策

为提高机场鸟击防范管理水平,实现探鸟雷达与多种驱鸟设备联动,提出一种基于支持向量机（SVM）的机场智能驱鸟决策方法。该方法包括训练和测试两部分。训练部分利用机场鸟类探测预警与驱赶联动系统获取的大量历史鸟情信息,结合专家知识,通过数据预处理与支持向量机训练,建立驱鸟策略分类模型;测试部分根据驱鸟实时智能决策结果,对驱鸟策略分类模型进行持续修正与优化。通过某机场的实测鸟情信息数据与若干驱鸟实例,证明驱鸟策略分类模型具有较高的决策正确率,并能够通过自身修正与优化应对各种新问题。本文方法针对实时鸟情信息,实现了多种驱鸟设备的优化组合,克服了驱鸟设备长期重复运行造成的鸟类对驱鸟设备的耐受性问题,极大改善了驱鸟效果。      

基于量子万有引力搜索的SVM自驾故障诊断

针对自动驾驶仪在实际测试过程中故障样本较少的情况,提出一种基于量子万有引力搜索算法(QGSA)的支持向量机(SVM)故障诊断模型。SVM能较好地解决小样本、非线性问题,适用于自动驾驶仪的故障诊断。为进一步提高万有引力搜索算法(GSA)对参数寻优的收敛速度和收敛精度,将基于GSA的QGSA应用于SVM的参数寻优中,以解决SVM由于参数选取不当导致过学习或欠学习的问题,从而获得最优的分类模型。通过模拟实验分析,当训练样本数量为50时,基于QGSA的SVM故障诊断模型分类准确率便能达到96.530 6%,而基于遗传算法(GA)的SVM故障诊断模型分类准确率为92.040 8%,基于GSA的SVM故障诊断模型分类准确率为91.632 7%。仿真实验结果表明,基于QGSA的SVM故障诊断模型具有更好的故障诊断能力。      

Optimization of One versus All-SVM using AdaBoost algorithm for rainfall classification and estimation from multispectral MSG data

In this paper, we implement the AdaBoost algorithm to optimize the classifications results of precipitations intensities carried out by One versus All strategy using Support Vector Machine (OvA-SVM). The model developed which combines the AdaBoost algorithm with a multiclass SVM is applied to images from the MSG (Meteosat Second Generation) satellite. Other variants to build multiclass SVMs, such as the OvO-SVM (One versus One SVM), SBT-SVM (Slant Binary Tree SVM) and DDAG-SVM (Decision Directed Acyclic Graph) are also implemented on which we tested the AdaBoost algorithm. The study showed that the AdaBoost algorithm performed better in the case of the OvA-SVM variant compared to the other variants.In order to evaluate the elaborated model, some classification techniques, such as the ECST Enhanced Convective Stratiform Technique (ECST), the SART where the Support vector machine, Artificial neural network and Random forest classifiers are combined, the Convective/Stratiform Rain Area Delineation Technique (CS-RADT) and the Random Forest technique (RFT) are applied. The classification results obtained show that AdaBoost with OvA-SVM (AdaOvA-SVM) presents very interesting performances where the evaluation parameters POD, POFD, FAR, BIAS, CSI and PC indicate the values 95.2%, 12.4%, 14.7%, 0.9, 88.1% and 96.5% respectively. Indeed, the AdaOvA-SVM technique has surpassed the CS-RADT, ECST and RFT techniques. As for the comparison with the SART, we noted that OvA-SVM presents very close results. The same trend was also observed when estimating precipitation. At the end of this study, it is shown that the AdaBoost algorithm performs better on a weak classifier or on a strong classifier operating in an unfavorable environment.     

Deep learning-based framework for monitoring of debris-covered glacier from remotely sensed images

In recent years, deep learning (DL) methods have proven their efficiency for various computer vision (CV) tasks such as image classification, natural language processing, and object detection. However, training a DL model is expensive in terms of both complexities of the network structure and the amount of labeled data needed. In addition, the imbalance among available labeled data for different classes of interest may also adversely affect the model accuracy. This paper addresses these issues using a new convolutional neural network (CNN) based architecture. The proposed network incorporates both spatial and spectral information that combines two sub-networks: spatial-CNN and spectral-CNN. The spectral-CNN extracts spectral information, while spatial-CNN captures spatial information. Moreover, to make the features more robust, a multiscale spatial CNN architecture is introduced using different kernels. The final feature vector is formed by concatenating the outputs obtained from both spatial-CNN and spectral-CNN. To address the data imbalance problem, a generative adversarial network (GAN) was used to generate data for the underrepresented class. Finally, relatively a shallower network architecture was used to reduce the number of parameters in the network and improve the processing speed. The proposed model was trained and tested on Senitel-2 images for the classification of the debris-covered glacier. The results showed that the proposed method is well-suited for mapping and monitoring debris-covered glaciers at a large scale with high classification accuracy. In addition, we compared the proposed method with conventional machine learning approaches, support vector machine (SVM), random forest (RF) and multilayer perceptron (MLP).     

基于记忆关联学习的小样本高光谱图像分类方法

高光谱图像（HSI）分类是遥感领域的基础应用之一。该任务旨在根据部分带类别标签的像素样本训练分类器，预测图像中剩余像素对应的类别标签。在实际应用中，由于人工标记样本成本过高，只能获得少量带标签的样本。针对少量样本无法准确描述数据分布从而导致训练过程过拟合的问题，提出一种基于记忆关联学习的小样本高光谱图像分类方法。考虑到无标签样本中包含大量与数据分布相关的信息，构建基于有标签样本记忆模块，并根据样本间的特征关联，利用不断更新的记忆模块学习无标签样本的潜在类别分布，构建无监督分类模型，并与传统的有监督分类模型进行联合学习。在多个高光谱图像分类数据集上的实验结果表明，所提方法能有效提升小样本高光谱图像分类的准确性。      

一种鲁棒性增强的机载网络流量分类方法

针对机载网络高度动态、高度不稳定造成流量监测设备难以在有限的监测周期内完成完整数据流负载特征的提取，限制了基于深度学习的流量分类方法的应用问题，提出了一种鲁棒性增强的机载网络流量分类方法。通过数据预处理及缺失样本处理方法将数据流映射为灰度矢量集合，基于完整的数据流训练数据集实现鲁棒性增强的长时递归卷积神经网络（RE-LRCN）分类模型的训练，在线上分类阶段，通过分类模型实现样本缺失数据流负载空间特征及数据流时序特征的提取，并进行数据流分类。通过在数据包缺失的流量测试数据集上的实验结果表明，所提方法可以有效抑制数据包缺失对分类准确性能的恶化。      

基于相控阵雷达波束篱笆的空间碎片数量与分布估计方法

随着载人航天与空间站等航天活动的增多,不能有效防护、也无法定期跟踪和编目的小尺寸(尤其是1～10 cm)碎片的危害越来越受到关注,这些碎片信息的获取依赖于统计采样技术.针对简化的相控阵雷达波束篱笆空间碎片探测模式,提出了一种采用统计技术估计空间碎片总数量以及高度和倾角分布的方法.将碎片穿越波束篱笆的过程用Poisson分布来建模,根据观测时段内穿越波束篱笆目标的平均到达率及测量的轨道高度和倾角数据来估计给定轨道高度范围或倾角范围内碎片的数量,进而得到碎片的总数量以及碎片数量随轨道高度或倾角的分布.在获取雷达散射截面信息时,该方法还可用于估计碎片数量随尺寸的分布.通过仿真实验验证了该方法的有效性.      

基于统计特征空间提取和支持向量机的极暗弱小天体检测方法

小天体检测是小天体防御和预警的前提。针对小天体目标信噪比低、检测难的问题,提出了基于统计特征空间提取和支持向量机（SVM）的极暗弱小天体检测方法。区别于传统方法基于时间或空间上目标的能量和背景噪声能量的瞬时能量差别或是瞬时能量差别的累积,对目标进行检测。该方法不依赖目标能量大小,提取运动目标穿过背景时对稳定性产生的扰动来反演运动目标。将输入的图像序列转化为单像元时序信号,划分时序窗口提取统计特征,关联形成统计特征空间,利用更高维度的变化特性检测目标变化。通过SVM将暗弱小天体检测问题转化为目标与背景的二分类问题,避开了较难解决的阈值分割问题同时具有更好的泛化性能。利用真实数据与其他经典方法进行对比分析,使得分类准确率提高4.02%。该方法能够适应更低的信噪比,在极低信噪比下仍表现出稳定的检测性能。     

Responses to satellite remote sensing opportunities in East and Southern Africa

Since 1978 the U.S. Agency for International Development (USAID) has funded a regional remote sensing project for East and Southern Africa. The project, hosted by the Regional Centre for Services in Surveying Mapping and Remote Sensing, has provided a programme of training courses, user services and project support. This included the equipping and establishment of a photo-laboratory complex for processing Landsat images and the provision of advice and support for agencies undertaking natural resources analysis.Response to the training programme has been very good. Courses are usually over subscribed and there is a continued demand for training. Assessments of the courses by participants are highly positive and the courses have featured consultants of international calibre. Requests for follow-up courses, and for specialist group training indicate a strong response to this training activity. User services are active, consultations with staff, use of the browse file and interpretation equipment and the purchase of data for project work all produce an average demand of 12 active enquiries per working week. The photo-laboratory is particularly active and demand for products exceeds available capacity.Project work is now being supported but limited resources restrict the range and amount of project activity. Response to the opportunities offered for projects has been favourable and this activity is ripe for expansion. The difficulty in expanding to meet the expressed demand is primarily financial. The east and southern Africa region is not economically strong and has a great need for natural resources data for development work and planning. The responses to satellite remote sensing opportunities will be limited by these financial constraints which effectively means by the level of international aid directed to this activity. For such aid to be effective it must be coordinated and firmly attached to the region. Such coordinated aid programmes would avoid fragmentation of distribution and applications work. Fragmentation could seriously disrupt the present growth in natural resources activity based on remote sensing data from satellites. The lack of financing for a regional ground receiving station is also a significant limiting factor.     

基于多分类AdaBoost的航空发动机故障诊断

对航空发动机运行数据进行数据挖掘的方法,是发动机故障诊断研究领域的重要研究内容。由于各种算法自身的局限性,通过某种单一算法很难大幅度提升故障分类的准确性。运用组合分类的AdaBoost算法,综合多个分类模型进行诊断,是提升故障识别精度的一种较好的方法。通过AdaBoost算法及其改进算法的结合,建立一种多分类的AdaBoost算法,以支持向量机（SVM）为基础分类器,进行综合诊断模型的建立。通过单位向量法、比值系数法和相关系数法将指印图中统计的故障标识数据进行处理,得到不受故障程度影响的训练数据,再进行建模。实验表明,AdaBoost相关结合算法能够显著提升分类器性能。根据实际故障案例,验证了所建立的诊断模型能够较好地用于发动机的故障诊断。      

基于变分推理的网络舆情传播模式分类

随着网络社交媒体的快速发展，对舆情信息的传播模式进行分析成为研究热点。针对网络舆情传播模式分类任务中，小样本数据多路径生成分类正确率低的问题，提出了舆情传播领域知识图谱结构定义，建立了基于微博数据的舆情传播知识图谱与舆情传播分析任务数据集，使用GraphDIVA模型进行舆情传播模式分类，并在自建数据集中进行了舆情传播模式分类25样本测试实验。结果表明:模型在经过20轮训练后，分类正确率从76%提升到89.4%，说明GraphDIVA模型在减少训练次数、提升分类正确率方面具有更优的效果。      

基于对抗和迁移学习的灾害天气卫星云图分类

针对卫星云图中的灾害天气数据存在严重不平衡问题，提出一个结合生成对抗学习（GAN）和迁移学习（TL）的卷积神经网络（CNN）框架以解决上述问题进而提高基于卫星云图的灾害天气分类精度。该框架主要包含基于GAN的数据均衡化模块和基于迁移学习的CNN分类模块。上述2个模块分别从数据和算法层面解决数据的类间不平衡问题，分别得到一个相对均衡的数据集和一个可在不同类别数据上提取相对均衡特征的分类模型，最终实现对卫星云图的分类，提高其中灾害天气的卫星云图类别分类准确率。与此同时所提方法在自建的大规模卫星云图数据上进行了测试，消融性和综合实验结果证明了所提数据均衡方法和迁移学习方法是有效的，且所提框架模型对各个灾害天气类别的分类精度都有显著提升。      

Assessment of the impact of dimensionality reduction methods on information classes and classifiers for hyperspectral image classification by multiple classifier system

Identification of the appropriate combination of classifier and dimensionality reduction method has been a recurring task for various hyperspectral image classification scenarios. Image classification by multiple classifier system has been evolving as a promising method for enhancing accuracy and reliability of image classification. Because of the diversity in generalization capabilities of various dimensionality reduction methods, the classifier optimal to the problem and hence the accuracy of image classification varies considerably. The impact of including multiple dimensionality reduction methods in the MCS architecture for the supervised classification of a hyperspectral image for land cover classification has been assessed in this study. Multi-source airborne hyperspectral images acquired over five different sites covering a range of land cover categories have been classified by a multiple classifier system and compared against the classification results obtained from support vector machines (SVM). The MCS offers acceptable classification results across the images or sites when there are multiple dimensionality reduction methods in addition to different classifiers. Apart from offering acceptable classification results, the MCS indicates about 5% increase in the overall accuracy when compared to the SVM classifier across the hyperspectral images and sites. Results indicate the presence of dimensionality reduction method specific empirical preferences by land cover categories for certain classifiers thereby demanding the design of MCS to support adaptive selection of classifiers and dimensionality reduction methods for hyperspectral image classification.     

航空液压泵磨损状况预测

磨损是航空液压泵典型的渐进性故障之一,因磨损量难以测量,对磨损状况进行准确的预测比较困难.针对上述问题,提出了基于多尺度数据的支持向量机预测方法,该方法将支持向量机用于时间序列预测的基本理论和数据多尺度分解、相空间重构方法结合,能更有效地挖掘时间序列的内在联系及变化规律.采用回油流量作为反映航空液压泵磨损状况的敏感信号,将其分解为趋势项和随机项,采用多尺度支持向量机作等维信息一步预测和多步预测,利用网格方法对预测模型参数寻优.对比传统支持向量机算法分析其预测精度,结果表明:多尺度支持向量机模型预测精度更高,适于中长期预测.      

非线性回归支持向量机的SMO算法改进

为了解决非线性数据和非线性函数的回归问题,采用了支持向量机序列最小优化算法.原始序列最小优化（SMO,Sequential Minimal Optimization）算法存在训练速度慢和训练结果不稳定的缺点,为了能加快SMO算法的训练速度和提高训练结果稳定性,通过改进优化乘子更新方法、采用双阈值法、预存核函数、增加停机准则等方法对SMO算法做了改进.仿真实验表明,改进的算法能很好地对非线性数据和非线性函数进行回归,具有比原始SMO算法更快的训练速度和稳定的训练结果.