Front Matter: Volume 8783

This paper investigates the segmentation algorithm proposed by Felzenszwalb and Huttenlocher¹ and its compatibility to 3D point clouds acquired with state-of-the-art 3D laser scanners. To use the algorithm, we adapt the range and intensity data to the smoothed graph structure used by the algorithm. We investigate the influence of the algorithm’s parameters to its performance and result that are meaningful to both the machines and the humans.

In this paper, we present a novel framework for multiple vehicles tracking in traffic crossroad that formulate multi-target tracking as an optimization problem. We set the optimizing decision model of multi-vehicles tracking based on characteristics of vehicles and traffic crossroad. In our formulation the problem of error propagation can be avoided through cutting down the error of detector by rejecting the improper detecting points during the optimizing process. Several challenging datasets are employed to validate the accuracy and robustness of our approach. A series of experiment results has demonstrated that our method is able to handle partial or even complete occlusions and can hardly be influenced by variant scale object.

Content-based image retrieval is an important task in many practical applications. All the existing algorithms can be divided into two categories, the global methods and the local methods. These methods both have their own drawbacks. Every kind of image property equally contributes to image retrieval applications in global methods, and only some selected ones are used in local methods. Generally, the global methods can not fully represent the specific properties of user interest, and many image properties are lost in local methods. In this paper, a novel method has been proposed, in which image properties are assigned with different weights, denoting the importance of the property in image retrieval application. The represented method simulates human perception of image foreground, and can make a trade-off between global methods and local methods. Experiments show that this method can improve image retrieval performance, without the lost of simplicity and compactness.

improve application of passive microwave imaging for object detection. In this study, we propose the FIPOCS (Fractal interpolation with Improved Projection onto Convex Sets) technique to enhance resolution. The experimental result shows that the resolution of passive microwave image is improved when utilizing the fractal interpolation to the LR image before applying the IPOCS technique.

Swarm intelligence-based image thresholding segmentation algorithms are playing an important role in the research field of image segmentation. In this paper, we briefly introduce the theories of four existing image segmentation algorithms based on swarm intelligence including fish swarm algorithm, artificial bee colony, bacteria foraging algorithm and particle swarm optimization. Then some image benchmarks are tested in order to show the differences of the segmentation accuracy, time consumption, convergence and robustness for Salt&Pepper noise and Gaussian noise of these four algorithms. Through these comparisons, this paper gives qualitative analyses for the performance variance of the four algorithms. The conclusions in this paper would give a significant guide for the actual image segmentation.

n this paper, we present an innovative iterative algorithm for tomographic reconstruction. Algebraic reconstruction technique (ART) which is considered as the core of iterative approach has been enhanced to ensure much finer and faster reconstruction. Backprojection has been customized to make it work even when the projections are not uniformly distributed. Contour information of the object has been combined with customized backprojection to ensure a better initial guess to start ART iterations. Based on experiments with both simulated and real medical images it has been shown that the proposed modality is capable of computing more accurate reconstructions in addition with lower computational cost than traditional ART.

With the popularity of digital camera and the application requirement of digitalized document images, using digital cameras to digitalize document images has become an irresistible trend. However, the warping of the document surface impacts on the quality of the Optical Character Recognition (OCR) system seriously. To improve the warped document image’s vision quality and the OCR rate, this paper proposed a warped document image correction method based on heterogeneous registration strategies. This method mosaics two warped images of the same document from different viewpoints. Firstly, two feature points are selected from one image. Then the two feature points are registered in the other image base on heterogeneous registration strategies. At last, image mosaics are done for the two images, and the best mosaiced image is selected by OCR recognition results. As a result, for the best mosaiced image, the distortions are mostly removed and the OCR results are improved markedly. Experimental results show that the proposed method can resolve the issue of warped document image correction more effectively.

In this paper, we propose a simple yet effective method for estimating defocus blur of microscopy images. The defocus blur at each edge point is estimated directly from local gradient profiles. To reliably detect edges and estimate blur in noisy and blur images, the image gradient is computed at multiple scales. The defocus blur at an edge point is firstly computed by the gradient ratio between different scales and is then recomputed by estimating the variance of the Gaussian gradient profiles. The proposed method is robust to noise and edge mis-localization. Experimental results on synthetic and real microscopy images demonstrate the effectiveness of our method.

In this paper, the spectral reflectance which ranges from 780nm to 1100nm for some typical targets is measured by using spectroradiometer. After the reflection properties analyzing, a near infrared (NIR) scene simulation method based on reflection properties of the typical target is presented. The experiment result shows the global features of NIR scene and laid the foundation of NIR scene simulation in future.

We introduce a new approach for salient motion in dynamic scenes. The proposed method is based on dynamic texture (DT) model and observability of the linear systems. Each video clip of τ frames is represented by DT in a holistic manner and the learned DT parameters are used to form an efficient formula of measuring observability. The formula is related to time-domain eigenvalues and eigenvectors. But the eigenvalue decomposition operation is not needed. The salient motion is detected by thresholding the observability values of each pixel location. Besides, we also introduce the system property of observability which can be used to reduce computational cost. Our method is tested on a challenging sequences set. Experimental results show that the proposed method has better trade-off between detection results and time efficiency than most current methods.

Face recognition algorithm based on compressed sensing and sparse representation is hotly argued in these years. The scheme of this algorithm increases recognition rate as well as anti-noise capability. However, the computational cost is expensive and has become a main restricting factor for real world applications. In this paper, we introduce a GPU-accelerated hybrid variant of face recognition algorithm named parallel face recognition algorithm (pFRA). We describe here how to carry out parallel optimization design to take full advantage of many-core structure of a GPU. The pFRA is tested and compared with several other implementations under different data sample size. Finally, Our pFRA, implemented with NVIDIA GPU and Computer Unified Device Architecture (CUDA) programming model, achieves a significant speedup over the traditional CPU implementations.

Besides illumination calculating during architecture and luminous environment design, to provide more varieties of photometric data, the paper presents combining relation between luminous environment design and SM light environment measuring system, which contains a set of experiment devices including light information collecting and processing modules, and can offer us various types of photometric data. During the research process, we introduced a simulation method for calibration, which mainly includes rebuilding experiment scenes in 3ds Max Design, calibrating this computer aid design software in simulated environment under conditions of various typical light sources, and fitting the exposure curves of rendered images. As analytical research went on, the operation sequence and points for attention during the simulated calibration were concluded, connections between Mental Ray renderer and SM light environment measuring system were established as well. From the paper, valuable reference conception for coordination between luminous environment design and SM light environment measuring system was pointed out.

In order to improve the invariance of the initial SIFT (Scale Invariant Feature Transform) and reduce mismatches when multiple local regions are similar, a novel method for image local invariant feature description based on SIFT is proposed. We introduce a new texture feature descriptor called center-symmetric improved local ternary patterns (CSILTP) that is a modified version of the well-known local binary patterns (LBP) feature and then fuse it with the global context information as the local feature descriptor in the SIFT algorithm. In the feature detection step, through iterative transformation, the initial feature point derived from the SIFT converges to affine invariant point and affine invariant region is extracted. In the feature description step, the dominant orientation is computed for each feature point. Then the image local feature descriptor is obtained by calculating the CS-ILTP and global context respectively. The feature matching experiment shows that the image local invariant feature descriptor present in this paper is invariant to image affine, scaling, rotation, illumination changes and so on. What's more, the number of correct matching features in artificial landmarks matching has increased by nearly 91% compared with the SIFT algorithm. The algorithm with high robustness and uniqueness is useful for the fields of robot navigation, image retrieval, etc.

Auto frame is a very large workpiece, with length up to 12 meters and width up to 2 meters, and it’s very easy to know that it’s inconvenient and not automatic to measure such a large workpiece by independent manual operation. In this paper we propose a new approach to reconstruct the 3D model of the large workpiece, especially the auto truck frame, based on multiple pulsed lasers, for the purpose of measuring the form and position errors. In a concerned area, it just needs one high-speed camera and two lasers. It is a fast, high-precision and economical approach.

Traditional non-blind image deblurring algorithms always use maximum a posterior(MAP). MAP estimates involving natural image priors can reduce the ripples effectively in contrast to maximum likelihood(ML). However, they have been found lacking in terms of restoration performance. Based on this issue, we utilize MAP with KL penalty to replace traditional MAP. We develop an image reconstruction algorithm that minimizes the KL divergence between the reference distribution and the prior distribution. The approximate KL penalty can restrain over-smooth caused by MAP. We use three groups of images and Harris corner detection to prove our method. The experimental results show that our algorithm of non-blind image restoration can effectively reduce the ringing effect and exhibit the state-of-the-art deblurring results.

Based on conventional camera calibration methods, a flexible approach for misaligned catadioptric camera from planar conic was presented. The catadioptric camera composed of a perspective camera and a hyperboloid mirror. The projection model of misaligned catadioptric camera was built and the projection functions were derived. Having known camera parameters, this method only requires the camera to observe a model plane which contains three concentric conics on the back of the revolution mirror. Then the mirror posture and attitude relative to the camera can be calculated linearly. The center of the concentric conics is on the mirror axis. This technique is easy to use and all computations are matrix operations in the linear algebra. The closed-form solution can be obtained without nonlinear optimization. Experiments are conducted on real images to evaluate the correctness and the feasibility of this method.

Sliding mode control (SMC) is a special nonlinear control method which have quick response, insensitive to parameters variation and disturbance, online identification for plants are not needed, its very suitable for nonlinear system control, but in reality usage, the chattering reduction and elimination is key problem in SMC. A sliding mode controller based on RBF neural network was designed based on RBF networks’ advantages and learning algorithm, the sliding mode controller was realized with adaptive law and the stability of the proposed control scheme is proved by Lyapnouv theorem. Simulation studies show the methods are effective and can applied into linear or nonlinear control system.

A novel method to measure the fiber length based on optoelectronic oscillator is firstly proposed in the experiment, showing that an optoelectronic oscillator can be used for measuring the fiber length by detecting the space of the oscillation frequency. The shorter the fiber is, the more precisely it can be measured.

In this paper, we present a simple and effective method to systematically derive an ensemble of Statistical Machine Translation (SMT) systems from one baseline linear SMT model for use in system combination. Each system in the resulting ensemble is based on a feature set derived from the features of the baseline model (typically a subset of it). We will discuss the principles to determine the feature sets for derived systems, and present in detail the system combination model used in our work. Evaluation is performed on the data sets for NIST 2004 and NIST 2005 Chinese-to-English SMT tasks. Experimental results show that our method can bring significant improvements to baseline systems with state-of-the-art performance.

After analyzing the working principle of the four-aspect fixed autoblock system, an energy-saving control model was created based on the dynamics equations of the trains in order to study the energy-saving optimal control strategy of trains in a following operation. Besides the safety and punctuality, the main aims of the model were the energy consumption and the time error. Based on this model, the static and dynamic speed restraints under a four-aspect fixed autoblock system were put forward. The multi-dimension parallel genetic algorithm (GA) and the external punishment function were adopted to solve this problem. By using the real number coding and the strategy of ramps divided into three parts, the convergence of GA was speeded up and the length of chromosomes was shortened. A vector of Gaussian random disturbance with zero mean was superposed to the mutation operator. The simulation result showed that the method could reduce the energy consumption effectively based on safety and punctuality.

In this paper, a new graph diffusion method is presented to improve the high-level feature extraction performance. In this method, we construct a semantic graph by describe the concepts as nodes and the concept affinities as the weights of edges, then we use the training set and its corresponding label matrix to estimate the concept relationship, where the relationship of two concepts were measured by the inner product of its corresponding row vector. We test the method on the high-level feature extraction task of TRECVID 2009 and the experimental results show the effectiveness of the method.

Identification (ID) systems utilizing visible light are an interesting topic, and have been attracting the growing attention of researchers recently. However, LED transmitters with a unique ID illuminate a limited coverage area, therefore, limiting the communication coverage. In order to extend the communication coverage, we proposed wireless mesh networks (WMNs) for relaying the information. Routing operations in WMNs might be challenging due to various factors. In this paper we analyze the link error probability and optical power loss. Based on our proposed network topology, we further found an effective scheme, multi-stage parallel interference cancellation (PIC), to alleviate the negative impacts due to interference which comes from multi-user data transmission.

Based on the NaSch Model, a new three-line cellular automata model emphasizing toll station on the high ways is built to discuss the impact of different amount of toll stations on the traffic flow. The models are as follows: Firstly, the process of cars driving is simulated. Secondly, the process of pulling station is simulated. In this part, two Cellular Automata Models are built separately for two cases, three tollbooths in the toll station and four tollbooths. The result shows that when the density of cars is on medium level, comparing with the toll station with three tollbooths, the toll station with four tollbooths can remit the traffic congestion effectively, but when the density of cars is too high or too low, the toll station with three tollbooths can do better.

In the process of high-speed magnetic flux leakage inspection, a large amount of signals containing pipeline status, characteristics and defects need to be collected and transmitted. High sampling frequency, high real-time, high precision and intellectualization for data collecting are long-term trend. This paper proposed a new design method of high speed data collecting system. In the system, dual CPU structure is applied in main controller and the high-speed communication between DSP and MCU is realized through dual-port RAM technology. With the Dual CPU structure being applied, the measurement, control and the communication are charged by two CPUs respectively, which can improve the system adaptability. Dual-port RAM can raise communication efficiency by improving the parallel processing ability among different type of multiprocessors.

Etching mechanisms and characteristics of bismuth zinc niobate (BZN) thin films were investigated in inductively coupled SF₆/Ar plasmas. The influences of various etching parameters including the gas flow ratio, process pressure, and ICP power on the etching results were analyzed. It is found that the chemical etching with F radicals was more effective than the physical sputtering etching with Ar ions for the inductively coupled plasma etching of BZN thin films. The mechanism of ion assisted chemical etching of BZN thin films in SF₆/Ar plasmas was proposed. A maximum etch rate of approximately 43.15 nm/min for the BZN thin film was obtained at the optimum etching conditions: 3/2 for the SF₆/Ar gas flow ratio, 10 mTorr for the process pressure, and 600 W for the ICP power. The surface morphology of the etched BZN thin film was observed, where was smooth and clean and no post-etch residues were remained.

The traditional multilevel model assumed independence between groups. However, the datasets grouped by geographical units often has spatial dependence. The individual is influenced not only by its region but also by the adjacent regions, and level-2 residual distribution assumption of traditional multilevel model is violated. In order to deal with such spatial multilevel data, we introduce spatial statistics and spatial econometric models into multilevel model, and apply spatial parameters and adjacency matrix in traditional level-2 model to reflect the spatial autocorrelation. Spatial lag model express spatial effects. We build spatial multilevel model which consider both multilevel thinking and spatial correlation.

Three-axis magnetoresistive sensor based on Anisotropic Magnetoresistive Technology(AMR) has excellent performance on detecting traffic Information. Data results for some features of microscopic magnetic field changes were obtained for road traffic after designing the test equipment of three-axis magnetic sensor and some experiments on road vehicles. The rules between the feature of microscopic magnetic field changes on road traffic and the impact by vehicles are found to realize the traffic Information detector for providing accurate data for intelligent transportation systems(ITS).

This paper gives the sufficient conditions of the blow-up of the solution for a class of nonlinear hyperbolic equation with initial boundary conditions infinite time and proves the existence and uniqueness of the local solution of the problem.

Chinese word segmentation technology is widely used in search engine domain. The architecture of Chinese word segmentation machine is designed with forward maximum matching algorithm and an improved double hashtable dictionary structure combining with the traditional participle dictionary mechanisms. The system includes Chinese information preprocess, dictionary management and word segmentation unit. The retrieval efficiency of this Chinese word segmentation machine is high and the process-speed is fast.

As for the problem of quality evaluation method of sea surface temperature (SST) observed by satellite remote sensing. An analysis model of SST is proposed based on the combination of observations at different time, different places and with different techniques. According to this model, Kalman Filter and the principle of non-negative matrix factorization can be used to fuse the SST data in temporal and spatial dimension when the data absence occurs. Through which an accurate estimation of SST observations will be made. The experiment results with SST data obtained in East China Sea in 2006, showed that the model presented in this article can obviously improve the precision of SST data estimation, which can provide accurate reference for the quality evaluation of marine information.

PID and fuzzy PID controller are applied into the pitch control system. PID control has simple principle and its parameters setting are rather easy. Fuzzy control need not to establish the mathematical of the control system and has strong robustness. The advantages of fuzzy PID control are simple, easy in setting parameters and strong robustness. Fuzzy PID controller combines with closed-loop optimal fuzzy reasoning (COFR), which can effectively improve the robustness, when the robustness is special requirement. MATLAB software is used for simulations, results display that fuzzy PID controller which combines with COFR has better performances than PID controller when errors exist.

For the security issues in distributed computing applications, combined with Web Services management system design and implementation, based on the research of Web Services and its security technology, this paper designs a distributed computing test system based on Web Services, in detail describes its key technologies, which further validates the advantages of Web Services technology for the distributed computing in loosely coupled network environment.

A protocol that allows any two entities to negotiate a shared session key is commonly called a key exchange protocol. If the protocol provides a function to authenticate each other, we call the protocol authenticated key exchange protocol (AKE). Password authentication key exchange (PAKE) is the AKE protocol in which the two entities share a humanmemorable password. Most of current PAKE relies on the existence of a public key infrastructure, which sometime is impossible for a certain environments such as low computational device due to the computation overhead. In this paper, we propose password–based authenticated key exchange using smart card. Compared to previous PAKE, our protocol is more efficient because our protocol is based on ECC. Thereby, the proposed protocol can be well applied to low computation device.

The dynamic simulation of oil migration and accumulation is an important issue on the research of petroleum exploration, and it is a numerical simulation process with special requirement on the framework mesh of 3D geological models, which means that the mesh should have same geometry and topology relation near the intersected part of geological surfaces. In this paper, basing on the conforming Delaunay triangulation algorithm to construct mesh of individual geological stratum or fault, a novel link-Delaunay-triangulation method is presented to achieve the geometric and topological consistency in the intersected line between two surfaces, also with the analysis of termination of our algorithm. Finally, some examples of the geological framework mesh are provided and the experimental result proved that the algorithm's effectiveness in engineering practice.

SLG continuous powder surface modifying machines have been widely applied in the powder engineering of China presently. Based on the industrial SLG-3/600 type modifying machine, a three-dimensional CFD model is developed to make the numerical simulations of the inner gas-particle flow field, including the gas velocity field, pressure field, particle concentration field and moving trace. The simulation results indicate that the velocity field and pressure field as well as the particle concentration field are distributed non-uniformly in the SLG modifying machine. Besides, a shortcut phenomenon is observed during the moving of the particles. The numerical simulations can not only depict the flow characteristics inside the SLG modifying machine, but also present useful reference to the mechanism study and optimal design to further improve its structure and working parameters in the future.

In consideration of the heavy image data needing to be transmitted and processed in real-time, a high speed transmission system is built using VMIC network for the simulation system of the infrared seeker. After the transmission from the image-acquiring computer to the super-control computer, the image data has to be compressed. To achieve the requirement of the interest region with high quantity compression, the JPEG 2000 format is selected. The Modular design method is used to perform the folded algorithm of JPEG 2000.

For ship targets at sea, because of their complicated ship movement, the conventional ISAR imaging algorithm can not be satisfied the needs of ship imaging. On the basis of analyzing the echo’s Doppler frequency caused by the complicated ship movements, a new method based on adaptive optimal kernel time-frequency representation for ship imaging is proposed. The instantaneous ship images at different moment are obtained. Simulation and real radar data verify its effectiveness.

The concentration and types of SF₆ in Gas Insulated Switchgear (GIS) play a decisive role in the devices’ insulating property. A quantitative analysis of SF₆ and its decompositions can help to find the reason of fault. In order to find the concentration information of some special ramifications of SF₆ from the infrared spectrum of GIS’s gas, this paper proposes Particle Swarm Optimization combines with Support Vector Machine to analysis the insulating medium SF₆ and its ramifications quantitatively. This paper studies the spectrum of several ingredients that are mordant to the insulator instruments in the ramifications, such as HF and SO_2. The mixed spectrum is divided into 13 parts, and the area of every part is calculated. The centre of each part is the characteristic peaks, and contains 35 wave numbers both side. These areas are used as the inputs of Support Vector Machine; the outputs is volumes of the three gases. The Particle Swarm Optimization is used to train the Support Vector Machine. The experiment shows Support Vector Machine based on Particle Swarm Optimization is time saved and accurate, which has practical significance and application potentiality.

License plate recognition system(LPRS) is one of the most important parts of the intelligent transportation system(ITS),and the license plate location is the most important step of the LPRS,it derectly affects the performance of the character segmentation and recognition afterward.In this paper,an effective algorithm of lisence plate location is proposed.In this method,Firstly we obtain high frequency coefficient through 1-D discrete wavelet transform. Then we process the image with median filter, binarization and morphology operation.Finally,we label and record the connected regions. Then we can locate the candidate license plates according to the region information. Experiment proved that our method performs well in the long range and complex scenes,and performs well on robustness.

In this paper, based on the analysis of existing embedded real-time operating system, the architecture of an operating system is designed and implemented. The experimental results show that the design fully complies with the requirements of embedded real-time operating system, can achieve the purposes of reducing the complexity of embedded software design and improving the maintainability, reliability, flexibility. Therefore, this design program has high practical value.

Important information, such as textures or edge, is often lost or blurred in the process of traditional denoising. In this paper we present a well balanced reaction-diffusion equation combined new harmonic analysis to resolve this problem better. A mixed diffusivity term is used to control the diffusion process to smooth and maintain essential image characteristics. While in reaction terms we employ wave atoms and curvelets shrinkage to preserve and enhance textures and edges. Experiment results and comparisons show the good performance of the proposed model for texture and edge preserving in the process of denoising especially as the noise increase.

As 3G (3rd-generation) networks evolve worldwide, the rising demand for mobile video services and the enormous growth of video on the internet is creating major new revenue opportunities for mobile network operators and application developers. The text introduced a method of mobile video transmission based on J2ME, giving the method of video compressing, then describing the video compressing standard, and then describing the software design. The proposed mobile video method based on J2EE is a typical mobile multimedia application, which has a higher availability and a wide range of applications. The users can get the video through terminal devices such as phone.

Fingerprint recognition technology has become the most reliable biometric technology due to its uniqueness and invariance, which has been most convenient and most reliable technique for personal authentication. The development of Automated Fingerprint Identification System is an urgent need for modern information security. Meanwhile, fingerprint preprocessing algorithm of fingerprint recognition technology has played an important part in Automatic Fingerprint Identification System. This article introduces the general steps in the fingerprint recognition technology, namely the image input, preprocessing, feature recognition, and fingerprint image enhancement. As the key to fingerprint identification technology, fingerprint image enhancement affects the accuracy of the system. It focuses on the characteristics of the fingerprint image, Gabor filters algorithm for fingerprint image enhancement, the theoretical basis of Gabor filters, and demonstration of the filter. The enhancement algorithm for fingerprint image is in the windows XP platform with matlab.65 as a development tool for the demonstration. The result shows that the Gabor filter is effective in fingerprint image enhancement technology.

Digital Mine is an important research content in the mineral field, the core is tunnel. A new three-dimensional data model of tunnel has been put forward, semi-tunnel was as volume element to construct the three-dimensional tunnel. According to this volume element, five basic modeling elements were abstracted to form the semi-tunnel. On the basis of semi-tunnel, it can build a model of simple tunnel, complex tunnel, composite tunnel and three-dimensional tunnel network. Then, the paper summarized data structure by analyzing the topological relationship between these tunnels and five elements. In terms of above new data model, the complete underground mine tunnel network will be constructed based on data structure. Thereby it can promote the development of theory of Digital Mine.

With rapid development of automatic control and network techniques, network-based remote monitoring is becoming an investigating hotspot in the elevator industry. At the same time as the development of wireless communication technology, remote wireless monitoring technology is applied more and more widely in recent years. A variety of wireless detection equipment is entering various industrial enterprises, and has been widely used. At present, there are many defects of the traditional monitoring system, such as poor real-time, low reliability, low intelligence. Based on the analysis of the difficulties to monitor the mobile terminal, this paper firstly analyzes the GSM/GPRS technology, and then discusses a design of the remote monitoring system based on wireless communication. The architecture of the monitoring center is introduced in detail. It is characterized by user-friendly, easy operate, good real-time and easy to extend.

A blind separation algorithm based on blur detection is proposed for a type of permuted alias image with blur difference. In the model of permuted alias image, space domain is selected as characteristic domain. Difference images are obtained by implementing four-phase-difference method, followed by partitioned into sub-blocks to compute variance value of them. Blurry images are separated from the permuted alias image by detecting the blurry sub-blocks with threshold operation. The simulation results show that blurry images can be effectively separated from the permuted alias image without respect to the location and size of permuting image.

Digital camera imaging model are introduced, the nonlinear responses of trichromatic digital camera systems are Corrected, The spectral characterization methods are compared based on low dimension linear models and polynomials regression. Reconstruction results showed that spectral reconstruction accuracy will be significantly improved by polynomial regression with the increase in the number of polynomials.

Processing tomato early blight seriously affect the yield and quality of its.Determine the leaves spectrum of different disease severity level of processing tomato early blight.We take the sensitive bands of processing tomato early blight as support vector machine input vector.Through the genetic algorithm(GA) to optimize the parameters of SVM, We could recognize different disease severity level of processing tomato early blight.The result show:the sensitive bands of different disease severity levels of processing tomato early blight is 628-643nm and 689-692nm.The sensitive bands are as the GA and SVM input vector.We get the best penalty parameters is 0.129 and kernel function parameters is 3.479.We make classification training and testing by polynomial nuclear,radial basis function nuclear,Sigmoid nuclear.The best classification model is the radial basis function nuclear of SVM. Training accuracy is 84.615%,Testing accuracy is 80.681%.It is combined GA and SVM to achieve multi-classification of processing tomato early blight.It is provided the technical support of prediction processing tomato early blight occurrence, development and diffusion rule in large areas.

Based on the analysis of particle filter algorithm, two improved mechanism are studied so as to improve the performance of particle filter. Firstly, hybrid proposal distribution with annealing parameter is studied in order to use current information of the latest observed measurement to optimize particle filter. Then, resampling step in particle filter is improved by two methods which are based on partial stratified resampling (PSR). One is that it uses the optimal idea to improve the weights after implementing PSR, and the other is that it uses the optimal idea to improve the weights before implementing PSR and uses adaptive mutation operation for all particles so as to assure the diversity of particle sets after PSR. At last, the simulations based on single object tracking are implemented, and the performances of the improved mechanism for particle filter are estimated.

With the inherent deficiency analysis of particle filter algorithm, proposal distribution with adaptive choice mechanism is studied. The adaptive mechanisms for proposal distribution include adaptive proposal distribution revised by the information derived from step-by-step Monte Carlo samples, Gaussian approximation adaptive proposal distribution, shrinking / growing adaptive proposal distribution, adaptive proposal distribution combined with other methods. At last, the simulations based on single object tracking are implemented, and the performance of the particle filter with adaptive proposal distribution is verified.

In the field of computer vision, image match is one of the most important research topics. It is used to match two or more images, for example, at different time, from different sensors or from different viewpoints. In this paper, we present a robust method for realtime image registration. First, the feature points of two images were obtained by using Adaptive Harris detector, and the matching points of the two images are determined by using global search. And then the maximum correlation principle is involved to adaptively determine the optimal affine transformation parameters. In the stage of real-time video image registration, we can determine the searching area for the related feature points in the current frame using the affine transformation matrix between the reference image and the previous frame, which enables the matching method effective. The experimental results showed that the presented method achieved 25fps, which is fast enough, and the registration errors are small enough demonstrating the high precision of the proposed registration method.

In order to quickly and exactly detect the subpixel image edge of micro-part, a fast subpixel edge detection method based on the property of Gaussian blurred edge model is proposed. Firstly, the approximate positioning of edge point was extracted by double-threshold segmentation; secondly a Zernike moment operator with a mask size of 5×5 was used to get rid of false edge points and relocate the edge with subpixel accuracy. Experiment results show that the subpixel accuracy and the running time of the method are 0.16 pixel and 0.94s. Therefore, the method is suitable for online threedimensional size detection of micro-part.

A moving object detection algorithm with sparse motion field estimation, motion classification and pixel-wise segmentation is proposed. Firstly, sparse motion field is recovered by fast corner detection and tracking. The corners that belong to the same motion pattern are classified according to their motion consistency, then, the resulting corner group is used to reconstructed scene image, and the foreground corners are identified by getting rid of the group with the least reconstruction error. Finally, optimal dense segmentation of the foreground is performed by using graph cuts, the energy function of which integrates corner motion, local color distribution and image edges. The proposed method is tested on the dataset of real complex scenarios and its effectiveness is demonstrated in the results.

Using threshold switch filter we can remove noise effectively with setting a threshold and dividing the proportion into different segments by this threshold, and then choosing a different filtering algorithm in each segment.Experiment results show that threshold switch filter is the best algorithm camparing to median filter, mean filter and alpha mean filter.This filter can preserve image details and remove most of noise, even in the case of high-density noise and small window.

In wind tunnel tests, the pose of the aircraft model in wire-driven parallel suspension system (WDPSS) is determined by driving several wires. Pose measurement is very important for the study of WDPSS. Using machine vision technology, Monocular Vision Measurement System has been constructed to estimate the pose of the aircraft model by applying a camera calibration, by extracting corresponding control points for the aircraft model, and by applying several homogeneous transformations. This article describes the programs of the measurement system, measurement principle and data processing methods which is based on HALCON to achieve the Solution of the pose of aircraft model. Through experiments, practical feasibility of the system is validated.

Microwave images are always blurred and distorted. Super-resolution analysis is crucial in microwave image processing. In this paper, we propose the WFIPOCS algorithm, which represents the wavelet-based fractal interpolation incorporates the improved projection onto convex sets (IPOCS) technique. Firstly, we apply down sampling and wiener filtering to a low resolution (LR) microwave image. Then, the wavelet-based fractal interpolation is applied to preprocess the LR image. Finally, the IPOCS technique is applied to solve the problems arisen by interpolation and to approach a high resolution (HR) image. The experimental results indicate that the WFIPOCS algorithm improves spatial resolution of microwave images.

Nowadays, different applications like automatic video indexing, keyword based video search and TV commercials can be developed by detecting and recognizing the billboard trademark. We propose a hierarchical solution for real-time billboard trademark recognition in various sports video, billboard frames are detected in the first level, fuzzy decision tree with easily-computing features are employed to accelerate the process, while in the second level, color and regional SIFT features are combined for the first time to describe the appearance of trademarks, and the shared nearest neighbor (SNN) clustering with ^x2 distance is utilized instead of traditional K-means clustering to construct the SIFT vocabulary, at last, Latent Semantic Analysis (LSA) based SIFT vocabulary matching is performed on the template trademark and the candidate regions in billboard frame. The preliminary experiments demonstrate the effectiveness of the hierarchical solution, and real time constraints are also met by our solution.

According to the needs of actual projects, one tiny lateral displacement detection method is introduced. This method is based on that the image of the moving object is analyzed and processed and image correlation matching is practiced, thus tiny displacement value of the objects is accurately calculated. Through the experiments, this method is proved to be of high accuracy, of strong anti-interference ability, safe and of convenient operation advantages and it can satisfy the use of certain special occasions.

Today many bug detecting tools use static program analysis techniques to discover the vulnerabilities in programs, and many static program analyses can be converted to CFL (Context-Free-Language) reachability problems, most of which are Dyck-CFL reachability problems, a particular class of CFL reachability problems based on Dyck languages. In order to speed up the static analyses formulated using the Dyck-CFL reachability problems, we propose an efficient algorithm of O(n) time for the Dyck-CFL reachability problem when the graph considered is a bidirected tree with specific constraints, while a naïve algorithm runs in O(n2) time. This is done by the bidirected-tree merging and an efficient method to determine the existence of the directed-path from the source to the destination.

Exploring efficient, reliable test case design methods has been tester pursuit of the goal. Along with the aerospace software logic complexity of improving and software scale enlarging, this requirement also gets more compelling. Test case design techniques suited for MC/DC improved test case design efficiency, increase the test coverage. It is suitable to test the software that logical relationship is complicated comparatively. Some software test tools provide the function to calculate the test coverage. And it can assess the test cases whether on the MC/DC or not. But the software tester needs the reverse thinking. The paper puts forward that design the test case by Unique-cause and Masking approach. And it proposes automatic generation method of test case on MC/DC. It improved the efficiency and correctness of generation the test case set on DC/DC.

Cooperation diversity allows mobiles share their antennas to achieve uplink transmit diversity. A particularly powerful variation of user cooperation framework is coded cooperation, which partitions the code words of each mobile and transmits portions of each code word through independent fading channels. This paper mainly focuses on coded cooperation, and its improved framework, which is called space-time cooperation, and investigates the application of turbo codes in both frameworks. The analysis and simulation results show that both of the frameworks achieve impressive gains, and space-time cooperation improves the performance of coded cooperation in fast fading.

Fresnel Seismic Tomography which uses a huge amount of seismic data is an efficient methodology of researching three-dimensional structure of earth. However, in practical application, it confronts with two key challenges of enormous data volume and huge computation. It is difficult to accomplish computation tasks under normal operating environment and computation strategies. In this paper, a Job-By-Application parallel computation strategy, which uses MPI (Message Passing Interface) and Pthread hybrid programming models based on the cluster, is designed to implement Fresnel seismic tomography, this method can solve the problem of allocating tasks dynamically, improve the load balancing and scalability of the system effectively; and we adopted the cached I/O strategy to accommodate the limited memory resources. Experimental results demonstrated that the program implemented on these strategies could completed the actual job within the idea time, the running of the program was stable, achieved load balancing, showed a good speedup and could adapt to the hardware environment of insufficient memory.

The detection of small targets that submerges in the homogeneous image background is a difficult problem in the design of small targets detection algorithm. The concept of local reverse entropy combining the concepts of local entropy with reverse entropy is proposed in this paper, which is to solve that problem by enhancing small targets. The norms of local signal-to-background ratio and elapsed time are used to demonstrate the performance of local reverse entropy map. Both quantitative analysis and qualitative comparison confirm the validity and efficiency of the presented approach.

The problems of detection, identification and tracking of moving target under the condition of moving camera are quite difficult in the application of computer vision. This paper proposes two different methods to the detection of moving target in the light of the size of the object. One approach integrates the image-interpolation technique and the active contour model, the other unites the image-interpolation technique, the modified local entropy algorithm and the region growing skill. The introduction of the image-interpolation technique is to estimate the global motion, i.e. background motion. The active contour model or the modified local entropy algorithm and region growing skill act upon the difference image, and then detect the moving target. It is demonstrated the reasonableness and efficiency of two detection approaches through experiments.

The acoustic echo path is usually very long and ranges from several hundreds to few thousands of taps. Frequency domain adaptive filter provides a solution to acoustic echo cancellation by means of resulting a significant reduction in the computational burden. In this paper, fast BLMS (Block Least-Mean-Square) algorithm in frequency domain is realized by using fast FFT technology. The adaptation of filter parameters is actually performed in the frequency domain. The proposed algorithm can ensure convergence with high speed and reduce computational complexity. Simulation results indicate that the algorithm demonstrates good performance for acoustic echo cancellation in communication systems.

Distributed source coding was proved to be suitable to correlated sources environments. This paper intends to explore a new way to improve distributed arithmetic coding and apply it in image coding. The proposed method models the correlation between sources as a Hidden Markov Model process. The decoder uses Viterbi algorithm to get a better error rate. Adjacent lines of the image are regarded as correlated sources in image coding process. Experiments that compress synthetic and real image data are carried on and the result shows that Hidden Markov Model-based distributed arithmetic coding can get better code rate and lower frame error rate.

In order to detect the errors in the software as much as possible and produce the software products with high quality, it is particularly important to do the quality evaluation of software testing work. In this paper the author did research of measurement of effectiveness of software testing from three aspects of testing method, testing process and testing result. Summarized the present technology condition of measurement of effectiveness of software testing at home and abroad, improved model of measurement of process of effectiveness of software testing, did the experiment of white box testing coverage rate and got the rough measurement value. The use of the proposed measure the effectiveness of software testing methods, analysis of data collected metrics, and then proposed to improve software testing techniques and software testing methods, so that make software testing effectively exert its positive function.

An efficient direction of arrival estimation method for wideband uncorrelated and coherent signals is proposed with a uniform linear array. Firstly, uncorrelated signals are resolved by applying a test of orthogonality of frequency subspaces method (TOFS). Secondly, at each frequency bin, the contributions of uncorrelated signals and noise are eliminated and a new matrix is formed. Then, forward spatial smoothing technique is performed on the new matrices and coherent signals are resolved with TOFS. The method can resolve more wideband signals than the array elements. Simulation results demonstrate the effectiveness and efficiency of the method.

A study on collaborative game in shared virtual environment with haptic feedback over computer networks is introduced in this paper. A collaborative task was used where the players located at remote sites and played the game together. The player can feel visual and haptic feedback in virtual environment compared to traditional networked multiplayer games. The experiment was desired in two conditions: visual feedback only and visual-haptic feedback. The goal of the experiment is to assess the impact of force feedback on collaborative task performance. Results indicate that haptic feedback is beneficial for performance enhancement for collaborative game in shared virtual environment. The outcomes of this research can have a powerful impact on the networked computer games.