Long-time Sequence Research Articles

Research on Wind Turbine Fault Detection Based on CNN-LSTM

With the wide application of wind energy as a clean energy source, to cope with the challenge of increasing maintenance difficulty brought about by the development of large-scale wind power equipment, it is crucial to monitor the operating status of wind turbines in real time and accurately identify the specific location of faults. In this study, a CNN-LSTM-based wind motor fault detection model is constructed for four types of typical faults, namely gearbox faults, electrical faults, yaw faults, and pitch faults of wind motors, combining CNN’s advantages of excelling in feature extraction and LSTM’s advantages of dealing with long-time sequence data, to achieve the simultaneous detection of multiple fault types. The accuracy of the CNN-LSTM-based wind turbine fault detection model reaches 90.06%, and optimal results are achieved for the effective discovery of yaw system faults, pitch system faults, and gearbox faults, obtaining 94.09%, 96.46%, and 97.39%, respectively. The CNN-LSTM wind turbine fault detection model proposed in this study improves the fault detection effect, avoids the further deterioration of faults, provides direction for preventive maintenance, reduces downtime loss due to restorative maintenance, and is essential for the sustainable use of wind turbines and maintenance of wind turbine service life, which helps to improve the operation and maintenance level of wind farms.

Electricity theft detection model based on the CEEMDAN-CNN-ViT

Abstract Most of the traditional power theft detection methods construct the model directly on the basis of the original power sequences, and do not simultaneously consider the long-period dependencies in the long-time sequences and the local connectivity features between periods, which limits the deep excavation of the behavioral laws of power users. In order to further improve the accuracy of power theft detection, this paper proposes a high-precision power theft detection model that integrates local anomaly filtering, energy consumption decomposition, and multi-feature fusion strategies. First, local anomaly filtering is used to eliminate local anomalies in the normalized energy consumption data to avoid the misleading effect of short-term abnormal behavior on the model. Then, the energy consumption decomposition based on CEEMDAN selects specific frequency band data that can accurately characterize the pattern of power theft users to improve the accuracy of power theft detection. Next, the long-time periodic features in the two-dimensional data and the short-time local features in the one-dimensional sequences are integrated by multi-feature fusion to enhance the adaptability of the model. The results show that the proposed model can effectively improve the detection accuracy, detection completeness, F1 score, and accuracy compared with the existing methods.

STGNets: A spatial–temporal graph neural network for energy consumption prediction in cement industrial manufacturing processes

Energy consumption is an essential indicator for conserving energy and reducing emissions in industrial manufacturing processes. However, the industrial production chain is a multi-variable, time-varying, nonlinear system, which poses significant challenges for energy consumption prediction, particularly under complex physical–chemical processes, even when the system is in a steady state. Furthermore, the poor accuracy cannot meet practical production requirements due to current prediction systems cannot integrate spatial–temporal information effectively. To solve this issue, this research provides a novel technique for final online goods quality prediction based on deep spatial–temporal graph neural networks (GNN). Our approach can capture hidden spatial information relationships and manage long-time sequences in the processing data by using a learnable dependency matrix and a stacked dilated convolution component. Moreover, these two primary components are organically merged into a single end-to-end optimization framework. The experimental results on the real-world test dataset demonstrate that our technique outperforms rival machine learning techniques. Our research exemplifies how graph deep learning architecture may be utilized to solve real-world problems in industries.

Deep Independent Recurrent Neural Network Technique for Modeling Transient Behavior of Nonlinear Circuits

This paper introduces a novel macromodeling method based on a recurrent neural network called deep independently recurrent neural network (DIRNN). The proposed method applies to time-domain modeling of nonlinear circuits and components, resulting in better training. It overcomes the vanishing and exploding gradient problems encountered with conventional RNNs. In conventional RNNs, all neurons in each layer are involved in recurrent connections that cause unnecessary connections, increasing the model’s complexity over time and making it hard to train for long-time sequences. To solve this problem, the proposed DIRNN’s neurons are independent of each other in recurrent connections, because each neuron only receives connections from its own previous hidden state. The validity of the proposed method is verified by modeling two nonlinear circuit examples, namely, a multi-stage driver terminated by a multi-line interconnect, and an on-chip voltage generator.

Multi-Stage Robust Dynamic Unit Commitment Based on Pre-Extended -Fast Robust Dual Dynamic Programming

To naturally characterize the long-time sequence coupled equipment such as energy storage systems (ESS) and to enforce the non-anticipativity of dispatch decisions under uncertain renewable energy, a multi-stage robust dynamic unit commitment (MRDUC) model is proposed in this paper. MRDUC transforms the multi-stage robust unit commitment into a dynamic programming form, which is the first attempt to combine day-ahead optimization with intra-day multi-stage robust dynamic dispatch. The unit commitment problem is solved in the pre-stage, and a simulation of intra-day multi-stage economic dispatch is done thereafter, which further refines the UC decision in an iterative way. A pre-extended -fast robust dual dynamic programming (PE-FRDDP) method is proposed to efficiently solve the MRDUC problem, which simultaneously optimizes the day-ahead UC and the intra-day decision policies. Case studies on the modified IEEE 6-Bus and 4 large-scale systems from 300 to 1888 buses demonstrate the effectiveness of the proposed model and the solution methodology.

Rural Development and Restructuring in Central China’s Rural Areas: A Case Study of Eco-Urban Agglomeration around Poyang Lake, China

This study aims to provide a scientific reference for rural reconstruction and revitalization in the areas covered by Eco-Urban Agglomeration Around Poyang Lake. Rural development and restructuring is a comprehensive process involving multiple elements and a long-time sequence. Accordingly, scientific knowledge concerning the evolution and characteristics of the spatial and temporal patterns of rural development and reconstruction is crucial for successively facilitating rural revitalization and ensuring the sustainable development of rural areas. In this study, a framework of rural development and restructuring was constructed for areas around Poyang Lake Eco-Urban Agglomeration based on the data regarding population, land, and industrial elements in the rural regional system, as well as the data of counties covered by Poyang Lake Eco-Urban Agglomeration. For this purpose, the entropy value and other research methods were used to analyze the level of rural development and the degree of rural reconstruction, as well as to identify the characteristics of rural reconstruction types. The study results revealed the following: (1) Rural Comprehensive Development Level has increased from 0.218 to 0.347, and the geographical development gap of the countryside has narrowed; however, the development level of each region and each factor continues to remain uneven. (2) The results demonstrated a wave-like advancement in the Rural Comprehensive Restructuring Degree, with a decreased Rural Population Restructuring Degree, an increased Rural Industry Restructuring Degree, and a decreased Rural Land Restructuring Degree. (3) Rural restructuring in the study area can be divided into six zones according to the level of rural development and the degree of rural restructuring, with Type I and Type III being the main types. Based on the above results, this research proposes optimizations for different rural development and reconstruction type zones.

Chaotic Bi-LSTM and attention HLCO predictor-based quantum price level fuzzy logic trading system

There are various indicators, i.e., Relative Strength Index (RSI) [1] [2], Moving Average Convergence Divergence (MACD) [3] [4] [5] [6], Stochastic Oscillator [7] [8] applications, to determine market movements with buying and selling decisions in computational Finance, but they have drawbacks that induced discrepancies to match against the best trading times at fixed order-triggering boundaries and delay problems. For example, RSI [1] [2]’s 70 and 30 overbuy and oversell are fixed boundaries. Orders can only be triggered when RSI’s value exceeds one of these boundaries, its computation only considers past market condition prompting indicators like RSI to trigger orders with delay. In this paper, we proposed a method to reduce these problems with advanced AI technologies to generate indicators’ buy and sell signals in the best trading time. Recurrent Neural Network (RNN) [9] has outstanding performance to learn time-series data automatic with long-time sequences but its ordinary RNN units [10] [11] such as Long-Short-Term-Memory(LSTM)[12] are unable to decipher the relationships between time units called context. Hence, researchers have proposed an algorithm based on RNNs’ Attention Mechanism [13] [14] allowing RNNs to learn information such as chaotic attributes [15] [16] [17] [18] and Quantum properties [19] [20] [21] contained in time sequences. Chaos Theory [15] [16] and Quantum Finance Theory (QFT) [22] are also proposed to simulate these two features (or attributes?). Quantum Price Level (QPL) [22] [23] is one of the well-formed QFT models to simulate all possible vibration levels to locate price. The system used in this paper consists of two components 1) neural network to predict future data and solve indicators lagging problem, and 2) fuzzy logic to solve fixed order-triggering boundaries problem. Its system design has two main parts 1) Chaotic HLCO Predictor consists of LSTM, Lee-Oscillator and attention mechanism to predict the High, Low, Close and Open, 2) QPL-based Fuzzy Logic Trading Strategy to receive the result and trigger trading signals. This new proposed model has obtained significant results in backtesting previous data and outperformed other traditional indicators to facilitate investment decisions when market changes constantly. Codes are available at https://github.com/JarvisLee0423/ Chaotic Quantum Finance AI Predicted Trading System.

Mapping of small water bodies with integrated spatial information for time series images of optical remote sensing

Small water bodies and their temporal changes are, especially in urban areas, closely related to people’s daily life and they have an impact on the living environment. For small water bodies mapping with optical remote sensing images, it is challenging to establish a balance between reducing incorrect water detection and increasing the integrity of water extraction. For time-series application, the temporal variability of the spectral information is also challenging for the widely used threshold methods, which are frequently solely based on spectral analysis. In this work we propose a spatial information-integrated small water bodies mapping (SWM) method to achieve a complete and accurate extraction and temporal change monitoring of small water bodies in complex urban environments. Our strategy is to make use of the spatial contextual information to account for the indistinguishability of small water bodies in spectral information. The roughness of the water index is calculated to enhance the contrast between water bodies and other thematic classes eventually present in the imagery. The proposed SWM was applied to different water indexes with an automatic threshold determination. We tested the effectiveness of the proposed algorithm using Landsat and Sentinel-2 multispectral data from three different urban environments (Shanghai, Guangzhou, and Wuhan in China) which include a variety of river courses and lakes. Nantan Lake of China is selected as a representative experimental area to test the SWM method by generating the inter- and intra-year water results. Overall accuracy (OA), F1 score (F1), producer’s accuracy (PA), and user’s accuracy (UA) were used to quantitatively evaluate the accuracy of the algorithm. Compared with the four state-of-the-art water detection methods (supervised random forest classification, hierarchical clustering, multi-band threshold, and modified normalized difference water index (MNDWI)), the proposed SWM algorithm achieves better water extraction performance. Small water bodies are found to be extracted more completely and incorrect water extractions are alleviated. The overall accuracy of the SWM algorithm achieves an average of approx. 97% (OA) and 0.95 (F1). The long-time sequence (from 2007 to 2021) and the short-time interval (monthly results of the year 2020) of the water extraction results by the SWM agree well with the ground truth data. The mean absolute area deviations between the water body extraction results of the SWM algorithm and ground truth is significantly smaller than that of the Global Surface Water dataset developed by the European Commission's Joint Research Centre (3.5% vs 49.0% for annual results and 6.0% vs 35.0% for monthly results).

Encoder–Decoder-Based Velocity Prediction Modelling for Passenger Vehicles Coupled with Driving Pattern Recognition

To improve the performance of predictive energy management strategies for hybrid passenger vehicles, this paper proposes an Encoder–Decoder (ED)-based velocity prediction modelling system coupled with driving pattern recognition. Firstly, the driving pattern recognition (DPR) model is established by a K-means clustering algorithm and validated on test data; the driving patterns can be identified as urban, suburban, and highway. Then, by introducing the encoder–decoder structure, a DPR-ED model is designed, which enables the simultaneous input of multiple temporal features to further improve the prediction accuracy and stability. The results show that the root mean square error (RMSE) of the DPR-ED model on the validation set is 1.028 m/s for the long-time sequence prediction, which is 6.6% better than that of the multilayer perceptron (MLP) model. When the two models are applied to the test dataset, the proportion with a low error of 0.1~0.3 m/s is improved by 4% and the large-error proportion is filtered by the DPR-ED model. The DPR-ED model performs 5.2% better than the MLP model with respect to the average prediction accuracy. Meanwhile, the variance is decreased by 15.6%. This novel framework enables the processing of long-time sequences with multiple input dimensions, which improves the prediction accuracy under complicated driving patterns and enhances the generalization-related performance and robustness of the model.

Spatiotemporal Evolution Characteristics of Urbanization in the Xiamen Special Economic Zone Based on Nighttime-Light Data from 1992 to 2020

In China and elsewhere, urban expansion is directly related to the important issues of social development, economic development, and the sustainable development of the ecological environment. Traditional statistical methods based on administrative regions lack geospatial information, which makes it difficult to analyze and explore in detail the development status and spatial differences of cities. In real time, nighttime light (NTL) remote sensing can reveal the spatial expansion change information of urban built-up areas (UB) on different scales, thus allowing for the analysis of urban spatial patterns and variations in urban development. Based on the long-time sequence NTL data from 1992 to 2020, this work studies the Xiamen Special Economic Zone by using the vegetation-water-adjusted NTL urban index (VWANUI) to extract the urban built-up areas and study the UB expansion patterns, the migration of the urban center of gravity, and intra-city differences. The result is a qualitative and quantitative temporal and spatial evaluation of Xiamen’s economic development characteristics. The results show that the UB of Xiamen expanded 349.219 km2 from 1995 to 2020, mainly concentrated in the period 2005–2020, during which time 79.44% of the expansion of the whole study period occurred. Throughout the study period, the urban center of gravity of Xiamen city shifts 8757.15 m to the northeast at the rate of 350.29 m/year in the direction of 74.88° (the urban center of gravity shifted from the inner island to the outer island). The total brightness of nighttime lights in Xiamen is gradually increasing, indicating that the level of urban economic development continuously improved over the measurement period, that human social activities have strengthened, and that the cross-island development strategy has produced certain results. These results provide data that describe urban development and policy formulation in Xiamen.

格尔木长时间序列遥感生态指数变化监测

PDF HTML阅读 XML下载 导出引用 引用提醒 格尔木长时间序列遥感生态指数变化监测 DOI: 10.5846/stxb202104261102 作者: 作者单位: 作者简介: 通讯作者: 中图分类号: 基金项目: 第二次青藏高原科学考察（2019QZKK0608）；城市功能单元的遥感智能识别技术研究（41971310） Monitoring of changes in the ecological index of long-time sequence Remote Sensing in Golmud, Qinghai Province Author: Affiliation: Fund Project: The National Basic Research Program of China (973 Program) 摘要 | 图/表 | 访问统计 | 参考文献 | 相似文献 | 引证文献 | 资源附件 | 文章评论 摘要:格尔木作为青藏高原重要的交通枢纽城市，其生态环境的可持续发展关系到该地区的经济繁荣与社会稳定。利用长时间序列的研究在节点选择方面缺少指向性，通常采用等间隔划分的方法来确定研究时间节点。基于此，通过Google Earth Engine （GEE）平台计算了格尔木近30年（1990-2019年）的遥感生态指数（RSEI），并使用非参数统计方法——Mann-Kendall突变检验，科学的确定研究的时间节点，再进一步分析格尔木生态环境时空变化并探讨其影响因素。研究表明：（1） RSEI可以有效地评价格尔木生态环境质量，30年间RSEI均值从0.38下降至0.33，生态环境整体呈退化趋势；（2）由突变检验确定1990、2001、2006、2015、2019年为研究的时间节点。各年份RSEI以较差为主，占总面积的45%以上，主要分布在柴达木盆地北部。RSEI等级变化以下降为主，2001年下降最为明显，2015年有所减缓；（3）受气候与人类活动不同程度的影响，格尔木RSEI变化趋势以下降为主，占区域总面积的89.73%，广泛分布在柴达木地区和唐古拉山地区，受气候因素影响较大；而变化趋势上升的仅占3.24%，主要分布在格尔木人类活动密集的市区、乡镇以及柴达木盆地成矿带，受人类活动影响较大。 Abstract:As an important transportation hub city in the Qinghai-Tibet Plateau, Golmud's sustainable and healthy development of the ecological environments is closely related to the local economic prosperity and social stability of the related regions. Most studies on ecological change monitoring with long-time sequence remote sensing imageries often ignore the directivity in the selection of key time nodes among the long time series, and they usually take methods such as equal interval partition to select time nodes. Based on the above situation, we first calculated the 30 year sequence data of Remote Sensing Ecological Index (RSEI) of the city of Golmud on the Google Earth Engine (GEE) platform and we scientifically defined the time nodes of the study area by a nonparametric statistical method, the Mann-Kendall mutation test method, and then further analyzed the temporal and spatial variation characteristics of the ecological environment in the city of Golmud and in the end discussions were carried out to find out the main influencing driving factors. The results showed that, (1) the index of RSEI could be used to effectively evaluate the quality of ecological environments in the city of Golmud and similar areas. The average value of RSEI in the past 30 years was gradually changed from 0.38 to 0.33, and the overall changing trend of ecological environment was degradation; (2) through the Mann-Kendall mutation test, the year nodes of 1990, 2001, 2006, 2015 and 2019 were selected as the final time nodes of the study area. For the above five time nodes, the grades of the index of RSEI are mainly in the state of poor, accounting for more than 45% of the total area of Golmud, and the regions with the state of poor were mainly distributed in the northern part of the Qaidam Basin. The changing trend of the grade of the RSEI index was mainly decreased during the whole time series, which was most obvious in the time node of 2001 and the changing trend began to slow down in 2015; (3) affected by the climate and human activities at different degrees, the changing trend of the RSEI index in Golmud was mainly decreasing, which accounted for 89.73% of the total area, and which was widely distributed in the Qaidam and Tanggula Mountains areas, was greatly affected by climate factors; While areas with increasing trend took only 3.24% of the total area. The areas with increasing changing trend of the RSEI index were mainly distributed in the urban and town areas with intensive human activities in the city of Golmud, and the metallogenic belt of Qaidam Basin, which were greatly affected by human activities. 参考文献 相似文献 引证文献

Spatio-temporal changes of AOD in Xinjiang of China from 2000 to 2019: Which factor is more influential, natural factor or human factor?

Aerosol optical depth (AOD), which represents the optical attenuation, poses a major threat to the production activity, air quality, human health and regional sustainable development of arid and semi-arid areas. To some degree, AOD shows areal air pollution level and possesses obvious spatio-temporal characteristics. However, long-time sequences and detailed AOD information can not be provided due to currently limited monitoring technology. In this paper, a daily AOD product, MODIS-based Multi-angle Implementation of Atmospheric Correction (MAIAC), is deployed to analyze the spatio-temporal characteristics in Xinjiang Uygur Autonomous Region from 2000 to 2019. In addition, the importance of influencing factors for AOD is calculated through Random Forest (RF) Model and the propagation trajectories of pollutants are simulated through Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Model. Spatio distribution of AOD presents a tendency that AOD value in northern Xinjiang is low while the value in southern Xinjiang is high. Regions with high AOD values are mainly concentrated in Tarim Basin. AOD in southern Xinjiang is the highest, followed by that in eastern Xinjiang and AOD value in northern Xinjiang is the lowest. Seasonal variation of AOD is significant: Spring (0.309) > summer (0.200) > autumn (0.161) > winter (0.158). Average AOD value in Xinjiang is 0.196. AOD appears wavy from 2000 to 2014 with its low inflection point (0.157) appearing in 2005, and then increases, reaching its peak in 2014 (0.223). The obvious downward tendency after 2014 shows that the use of coal to natural gas (NG) conversion project improves the conditions of local environment. According to RF Model, NG contributes most to AOD. HYSPLIT Model reveals that aerosol in southern Xinjiang is related to the short-distant carriage of dust aerosol from the Taklimakan Desert. Aerosol there can affect Inner Mongolia through long-distant transport. Blocked by the Tianshan Mountains, fine dust particles can not cross the Tianshan Mountains to become a factor contributing to AOD in northern Xinjiang.

Sequential Mechanical Fault Diagnosis in High Voltage Circuit Breaker using Attention Mechanism

To address the shortcomings of traditional shallow-based model in terms of feature extraction and generalization capacity, this paper provides a high-voltage circuit breaker mechanical condition online detection scheme by using attention mechanism to locally weight the sample correlation and combining convolutional neural network (CNN) and long short-term memory (LSTM) network. The network uses convolutional layers for feature transformation of the raw vibration data, combined with the local time-domain feature representation capability of the gated unit, to extract fault-sensitive features. The temporal deep learning model using the attention mechanism enables it to extract global features of long-time sequences while making important features obtain higher weight parameters. The resulting model improves the overall learning and recognition efficiency. The experiments on the vacuum circuit breaker show that the proposed model has better performance than the classical diagnostic method based on the support vector machine.

Deep Learning of Complex Batch Process Data and Its Application on Quality Prediction

Batch process quality prediction is an important application in manufacturing and chemical industries. The complexity of batch processes is characterized by multiphase, nonlinearity, dynamics, and uneven durations so that modeling of these batch processes is rather difficult. Moreover, there are other challenges in the face of quality prediction. Specifically, the process trajectories over the whole running duration potentially make specific contributions to the final targets so that the prediction issue embraces tremendously high-dimensional inputs but very low-dimensional outputs. This means that the prediction suffers from a severe dimensional imbalance between inputs and outputs. Motivated by these difficulties, this paper proposes a new deep learning-based framework for complex feature representative and quality prediction. Long short-term memory (LSTM) is used to extract comprehensive quality-relevant hidden features from a long-time sequence in each phase, significantly reducing the predictor dimensions. And these features from different phases are further integrated and compressed by a stacked auto-encoder (SAE). A practical industrial example testifies to the efficacy of the proposed framework.

Quantitative Estimation of Biomass of Alpine Grasslands Using Hyperspectral Remote Sensing

In order to promote the application of hyperspectral remote sensing in the quantification of grassland areas’ physiological and biochemical parameters, based on the spectral characteristics of ground measurements, the dry AGB and multisensor satellite remote sensing data, including such methods as correlation analysis, scaling up, and regression analysis, were used to establish a multiscale remote sensing inversion model for the alpine grassland biomass. The feasibility and effectiveness of the model were verified by the remote sensing estimation of a time-space sequence biomass of a plateau grassland in northern Tibet. The results showed that, in the ground spectral characteristic parameters of the grassland’s biomass, the original wave bands of 550, 680, 860, and 900 nm, as well as their combination form, had a good correlation with biomass. Also, the remote sensing biomass estimation model established on the basis of the two spectral characteristics (VI2 and Normalized Difference Vegetation Index [NDVI]) had a high inversion accuracy and was easy to realize, with a fitting R2 of 0.869 and an F test value of 92.6. The biomass remote sensing estimate after scale transformation had a standard deviation of 53.9 kg/ha from the fitting model established by MODIS NDVI, and the estimation accuracy was 89%. Therefore, it displayed the ability to realize the estimation of large-scale and long-time sequence remote sensing biomass. The verification of the model’s accuracy, comparison of the existing research results of predecessors, and analysis of the regional development background demonstrated the effectiveness and feasibility of this method.

Damped Oscillations of the Energy of a Bosonic Bath due to Spectral Gaps and Special Initial Correlations

The energy of the bosonic bath and the flow of quantum information in local dephasing channels are studied over short and long times in case the distribution of frequency modes of the bosonic bath exhibits a low-frequency gap. The initial conditions consist in special correlations between the qubit and the bosonic bath or are factorized, and involve thermal states of the whole system or of the bath. The low-frequency gap generates damped oscillations of the bath energy around the asymptotic value, for the correlated initial conditions, and induces the open system to alternately loose and gain information, for the factorized initial configurations. The long-time oscillations of the bath energy become regular and the frequency of the oscillations coincides with the upper cut-off frequency of the spectral gap. Regular long-time sequences of intervals are found over which the bath energy increases (decreases), for the correlated initial conditions, and information is lost (gained) by the open system, for the factorized initial configurations, even at different temperatures. This relation is reversed, if compared to the one obtained without the low-frequency gap, and can fail if the spectral density is tailored near the spectral gap according to power laws with odd natural powers.

DUK-SVD: dynamic dictionary updating for sparse representation of a long-time remote sensing image sequence

Sparse representations of data or signals have drawn considerable attentions in the past decade. In this paper, we focus on the problem of training high-efficacy dictionaries for remote sensing images of massive long-time sequences. By extending the classical K-SVD, we propose a new dictionaries learning algorithm. Different from K-SVD, in the proposed incremental K-SVD algorithm, we selectively train a certain number of atoms when each new batch of sample data are added into the training process; current dictionary are replenished by the selected and enhanced atoms. The new atoms are initialized by information entropy. Meanwhile, we introduce an uncertainty metric to determine whether or not new atoms should be added into the current dictionary. To efficiently and sparsely represent the long-time sequence data set, we also de-correlate the dictionary based on new atoms by introducing a mutual coherence constraint into the atom updating stage. The method presented in this paper aims to adaptively and dynamically train the dictionary from big data. Two other state-of-the-art dictionary learning methods such as online dictionary learning (ODL) and recursive least squares dictionary learning algorithm (RLS-DLA) who also could train the dictionary using relatively large data, are comprehensively compared with the proposed algorithm in both sparse model and error model. In the sparse model, the reconstruction error of the DUK-SVD dictionary was smaller than ODL and RLS-DLA. In the error model, the sparsity of the DUK-SVD was higher than ODL and RLS-DLA. We can also observe that in the sparse model the proposed DUK-SVD often consume fewer computing time than ODL.

Robust tracking with per‐exemplar support vector machine

The authors extend exemplar representation to the field of tracking and propose a robust tracking algorithm with per-exemplar support vector machine (SVM) classifiers. First, the authors train the simple yet effective exemplar SVM classifier using the target object as the single positive and mining its surroundings as hard negatives. Second, the authors propose an online ensemble tracker, which integrates the useful ‘key historical templates’ of the target to refine the current template, leading to better discriminative power of tracker and effectively decreasing the risk of drift. Experiments on challenging sequences demonstrate that the tracker performs well in accuracy and robustness, especially under the sequences with strong illumination variation and scale variation, as well as pose change and partial occlusion in the long-time sequence.

Active-matting-based object tracking with color cues

In view of drifting issue in object tracking, the conventions are prone to degenerate due to the inaccuracy in appearance models. In this paper, we propose an active-matting-based visual tracker to give more precise contours in targets. The basic idea is to explore the biological inspired color surface coding theory to refine the original interest point detector, which benefits for robust representation to extract the suitable interior object areas for matting. In order to generate accurate pixel-wise labels of each frame for matting, we have both foreground and background interest points using k-d trees between two successive frames, under the similar geometric constraints from the object silhouette in the previous frame. The resulting tracker achieves performance competitive with the state-of-the-art in different color video sequences, especially under the scenarios with strong illuminations and posture variations, as well as small-scale targets in the long-time sequence.

Russian/Former Soviet Union Experience in Professional Small Telescope Usage

FSU astronomers traditionally use small telescopes (⊘ ≤1.5 m, hereafter STs) for both science and education. Russian/FSU experience here is among the largest world-wide. There are only 2 large and moderate-sized facilities in whole Russia: the 6 m telescope of SAO RAS and Russian-Ukrainian 2 m one on the 3100 m high peak Terskol in Central Caucasus.Equipped with good light receivers and handled by skilled observers, STs can produce first class scientific data. Important results are typically yielded by long-time sequences of observations and/or new observational “know how”: good instrument/receiver design, appropriate selection of objects and moments, etc. Examples of what has been done with STs in FSU, within my memory, (in the last ≃ 1/3 century) are listed below, without a list of references, because of lack of space. The author plans to publish a larger article on this subject in Astr.&Aph.Trans.