Search

Image super-resolution focuses on achieving the high-resolution version of single or multiple low-resolution images. In this paper, a novel super-resolution approach based on morphological component analysis (MCA) and dictionary learning is proposed in this paper. The approach can recover each hierarchical structure well for the reconstructed image. It is integrated mainly by the dictionary learning step and high-resolution image reconstruction step. In the first step, the high-resolution and low-resolution dictionary pairs are trained based on MCA and sparse representation. In the second step, the high-resolution image is reconstructed by the fusion between the high-resolution cartoon part and texture part. The cartoon is acquired by MCA from the interpolated source image. The texture is recovered by the dictionary pairs. Experiments show that the desired super-resolution results can be achieved by the approach based on MCA and dictionary learning.

Quantization is an essential but often ignored part of the realization of compressive sampling (CS), and the analysis of quantization noise arise from CS is incomplete and not sufficient until now. The quantization noise is generated from quantizing CS values by a uniform quantizer under ideal and noise conditions. And also, the auto correlation function and power spectrum have been derived. It is concluded that the quantization noise is always uncorrelated with the input signals, the quantization noise is white and the spectrum is white noise spectrum. On this basis, we analyze the reconstruction error introduced by quantization noise quantitatively and give the upper and lower bounds of reconstruction error. Simulation results validate the validity of the analysis for further.

In order to improve the accuracy and reliability of nonlinear state estimation problems with unknown but bounded noises in power system, an extend set-membership filter for state estimation in power systems is present in this article. The method is based on three sampling sine wave relational model. It overcomes the poor robustness, divergence and weak traceability of kalman filter, avoids the complex calculation process of traditional extend set-membership filter. Compared with kalman filter, the simulation results show that extend set-membership filter algorithm can track signals faster and more accurately.

Traffic flow detection plays an important role in Intelligent Transportation System (ITS). However, the conventional traffic flow detection approaches are high cost or complex installation. In this paper, a reliably vision-based traffic flow detection approach is proposed. In this approach, Gaussian mixture model (GMM) is employed to model the dynamic background of traffic scene. Then, the binary foreground contours are extracted by image subtraction. Comparing the binary vehicle contours’ location and the current frame, the real and complete vehicles are obtained for detecting and monitoring. In the part of vehicle counting, to gather the vehicle flow parameter in each lane of the road and avoid the trouble of counting vehicles repeatedly, a discriminative method is presented to classify vehicles into different lanes. Experiment shows that a desired result can be achieved in the traffic flow detection system by the vision-based approach.

Manifold learning is a popular recent approach to nonlinear dimensionality reduction. While conventional manifold learning methods are based on the assumption that the data distribution is uniform. They are hard to recover the manifold structure of data in low-dimension space when the data is distributed non-uniformly. This paper presents an improved Laplacian Eigenmaps algorithm, which improved the classical Laplacian Eigenmaps (LE) algorithm by introduce a novel neighbors selection method based on local density. This method can optimize the process of intrinsic structure discovery, and thus reducing the impact of data distribution variation. Several compared experiments between conventional manifold learning methods and improved LE are conducted on synthetic and real-world datasets. The experimental results demonstrate the effectiveness and robustness of our algorithm.

This paper proposed a novel two-stage hybrid container throughput forecasting model. Time series in reality exhibits both linear and nonlinear characteristics and individual models are not able to describe the two features simultaneously. Therefore, we combine linear model SARIMA (seasonal autoregressive integrated moving average) and nonlinear model ANN (artificial neural network). In order to break through the limitations of traditional hybrid models, based on the identified parameters of SARIMA in first stage, the structures of several ANN in second stage could be decided. Finally, we validate the proposed hybrid model 5 performs best with case study in Shanghai port.

This paper studies uncertain demand by a supplier and manufacturer two echelon closed-loop supply chain system model setting up and solving the problem, gets the manufacturers to control the transfer function of the z-domain expression and manufacturer System Model solving analysis. Sensitivity analysis of the model parameters, and comes to different parameters of the closed-loop supply chain model.

Storytelling activity is an effective method to enhance children's presentation ability, logical thinking and imagination in elementary education. In this paper, researchers observed elementary school children in the course of Performing Art and discovered the difficulty for them to operate the puppets. Hence, the purpose of this paper is to take advantages from both technology and storytelling activities to solve this kind of problems. A real-time chromakey technique system which is able to composites the actors and the scenes in real-time is conducted for children to create plays intuitively as well as record their performance for sharing, therefore reduce the frustration and increase the motivation of children in creating stories.

It is important to analyze and predict the measuring data trend in industrial measuring and controlling process. The paper introduces a method for predicting the trend of the current measuring data based on clustering the historical data. It calculates the similarities of the current trend and the bases result from the clustering. And with these similarities, the future trend of the current measuring data can be predicted , the combination of the above bases representing low frequency and a reviser representing high frequency. The simulation shows the weights of high or low frequency have effect on the precision of predict results. It is also found that the proposed method can predict more precisely than the RBFNNs method in high frequency.

This paper addresses a distributed model predictive control (DMPC) scheme for multi-agent systems with communication distance constraints. Firstly, the communication distance constraints are dealt as non-coupling constraints by using the time varying compatibility constraints and the assumed state trajectory. Obviously, the control performance for all system is influenced by the time-varying compatibility constraints. Secondly, the deviation punishment is involved in the local cost function of each agent to penalize the deviation of the computed state trajectory from the assumed state. The value of the time-varying compatibility constraints is set according to the deviation of previous sample time. The closed-loop stability is guaranteed with a large weight for deviation punishment. A numerical example is given to illustrate the effectiveness of the proposed scheme.

This paper presents a multi-cellular-ant algorithm for large scale capacitated vehicle routing problem with restrictive vehicle capability. The problem is divided into corresponding smaller ones by a decomposition methodology. Relative relationship between subsystems will be solved through cooperative performance among cellular ants to avoid trivial solutions. The empirical results composed with adaptive ant colony algorithm and traditional collaboration show more efficiency and availability.

In this paper an improved Mean-shift tracking algorithm based on adaptive multiple feature fusion is presented. A two-class variance ratio is employed to measure the discriminate between object and background. The multiple features are Fused by linear weighting to realise Mean-shift tracking using the discrimination. Furthermore, an adaptive model updating mechanism based on the likelihood of the features between successive frames is addressed to alleviate the mode drifts. Based on biology vision theory, colour, edge and texture cue are employed to implement the scheme. Experiments on several video sequences show the effectiveness of the proposed method.

A new modeling assumption for wireless sensor networks was put forward. Under the modeling assumptions, a new energy conservation scheme was designed and implemented for efficient data propagation. The network conditions were locally monitored, and accordingly the sleep-awake schedules of the nodes was adjusted to improve operation choices. The protocol was tested in software. Using the network simulator ns-2, the adaptive energy management protocol (AEMP) was compared with the Directed Diffusion Protocol (DDP) and the achieved performance was evaluate with a detailed simulation study. The advantages of our protocol are that it do not require knowledge of topology and not requires explicit coordination of sleep-awake schedules between nodes. Simulation results show that this new scheme is simple, and Directed Diffusion with AEMP outperforms again the stand-alone Directed Diffusion.

Artificial neural networks can be used as intelligent controllers to control non-linear, dynamic systems through learning, which can easily accommodate the non-linearities and time dependencies. However, they require large training time and large number of neurons to deal with complex problems. Taking benefit of the characteristics of a Generalized Neuron that requires much smaller training data and shorter training time, the pseudo-linear neural network (PNN) based model predictive approach used in the single and multi-machine power system studies is proposed in this paper. A simulation is carried out. It is demonstrated that the proposed control strategy is applicable to some of nonlinear systems.

The paper adopts rough sets theory and fuzzy ARTMAP method to explore the online adaptive contingency classification in power system transient stability control. On the basis of contingency vector space model, the rough sets theory is applied to generalize the information system comprised by contingency samples set, and compute the best reducing properties set. So dimension of contingency feature space is reduced greatly, and disturbance in contingency classification is decreased too, which improves the efficiency of classification. In addition, using the advantage of adaptive classification and incremental learning of Fuzzy ARTMAP neural network, the online adaptive classification of contingency is achieved.

As optical fiber is subject to faults, normal communication will be affected. An intelligent method of detection and location for communication optical fiber is put forward in this paper. According to spatial characters of geographic distributing of optical fiber network, nodes and links topo model of the network is built. Adopting the ANN algorithm in this paper, the nodes are classified according to the structure of optical fiber communication network, an effective ergodicity detection strategy of nodes and links is built, and free optical fiber and optical cable are detected termly. Through the simulation, the method which is put forward in this paper is validated. When optical fiber communication network extend, the method can form a new ergodique detection strategy of nodes and links based on the nodes classified, and the on-line dynamic detection of communication optical fiber can be realized.

Raster map should be corrected after scanned because of the errors caused by paper map deformation. In the paper, the deficiency of the polynomial fitting method is analyzed. The paper introduces an ANN (Artificial Neural Network) correcting method that utilizes the advantage of its function approximation ability. In the paper, two types of ANNs, BP and GRNN, are designed for the correcting. The comparing experiment is done with the same data by the polynomial fitting and ANN methods, utilizing the MALAB. The experiment results show that the ANN methods, especially the GRNN method, performances far better than the polynomial fitting method does.