Search

In the tension measurement by using magneto-elastic tension sensor, generally the changes of hysteresis loop with tension is applied. But due to the changes in environment temperature, the changes of hysteresis loop contain not only the tension information but also the temperature influence. In order to eliminate the influence of temperature change, a temperature compensation method used in rod-like structure tension measurement is proposed in this paper. The proposed method involves three key steps. First, a curve of hysteresis loop change (CHLC) is defined to reflect the influence of temperature and tension. It contains two components: 1) the tension component of CHLC and 2) the temperature component of CHLC which is an unknown nonlinear curve related to the temperature changes. Second, a prediction model of temperature component of CHLC based on the neural network is proposed. Finally, the temperature influence on CHLC is eliminated by calculating the difference between the CHLC and predicted temperature component of CHLC, and the predicted tension component of CHLC is obtained. The experimental results show that the CHLC can be used as the basis of the temperature compensation method. Moreover, the temperature influence can be analyzed by CHLC when the tension is invariable, whereas the tension influence can also be analyzed when the temperature is invariable. The temperature component of CHLC can be obtained rapidly and accurately by the neural network-based prediction model with a prediction error less than 10−5. The degree of CHLC distortion caused by temperature change is reduced from 10−2 to 10−5 with the proposed temperature compensation method.

Vegetation is a land-use type that attracting growing attention in urban studies. Extraction of urban vegetation information can be achieved by land-use/land-cover (LULC) classification. It has been proved that SAR remote sensing provides complementary information in LULC classification. However, how the complementary information from SAR features contribute to the classification in highly-dense urban area is still unclear. In this study, taking the highly urbanized area of Hong Kong as the study site, comprehensive studies were conducted to investigate the contribution of optical and polarimetric SAR features for LULC classification and urban vegetation extraction. Firstly, large amount of redundant information was found within optical and SAR features sets by Pearson correlation analysis. Then analysis based on J-S distance revealed that optical features have much stronger capabilities on LULC classification and some key features such as NDVI and Mean value of GLCM carry distinct information for discriminating certain land-use types. Supervised classification based on random forest further proved this result. The complementary mechanism of optical and polarimetric SAR features can be clearly observed in the radar chart. Therefore, when fusing optical and SAR features on the feature-level, the classification accuracy further improved. It was discovered that the volume scatter mechanism plays a vital role in the feature-level classification model. Finally, the overall classification accuracy of the classification model does not significantly grow after several key optical and SAR features are inputted, confirming again the high correlation between the features. In vegetation extraction, optical features provide most information and achieved the high accuracy of 99.02%. The results discovered in the above research could benefit the designing algorithms and selecting features for high accuracy urban LULC classification and vegetation extraction based on optical and SAR data.

To achieve powerful infringement retrieval for reference images in digital publications, a new improved Scale Invariant Feature Transformation (SIFT) algorithm has been proposed in this paper. The retrieval process in the improved algorithm is innovatively divided into two stages to achieve coarse retrieval and fine retrieval respectively. In coarse retrieval, Geographical Statistics (GeoStat) is creatively used to describe the global spatial relationship of key-points in different orientations in an image and then generate a 144-dimensional feature vector to represent each image. In fine retrieval, only partial images, which are highly similar to the query image obtained from results of coarse retrieval, need to be considered. And the indexing and matching process in the improved algorithm is improved by adding a judgment process to improve the matching speed and reduce the mistaken matching rate. Experimental results show that the proposed algorithm has more advantages in retrieval speed and higher retrieval accuracy than the original SIFT algorithm. And the proposed algorithm is also more suitable for infringement retrieval of reference images in digital publications than the original one.

In recent years, non-destructive evaluation (NDE) has attracted increasing attentions. Different electromagnetic signals have been used for studying the microstructure of materials. However, the features extracted for NDE share the limitation that these features were separately designed and roughly assembled. In order to overcome this limitation, a group of features that include the basic binary pattern of eddy current (BPE) and three improved features named simple-BPE, spat-BPE, and direction BPE (DBPE) are proposed to predict the residual strain (RS). Four features are designed through encoding the modulus variation of multi-frequency, multi-spatial, and multi-direction eddy current signal. In the experiment part, two types of experiments are used to test our four features. Based on the results, the performance of BPE-based methods outperforms the conventional features. Within four proposed features, the best prediction accuracy (0.03057997 for C45E4 and 0.03167161 for AlCu4Mg1) is obtained by adopting spat-BPE; the most efficient feature is simple-BPE, and a balanced performance is achieved by BPE and DBPE, performed not as good as the other three features. It has been proved that our binary-encoding-based features are beneficial to building the relationship between the NDE signal and the property of material (e.g., RS) and could be applied to a wide range of applications.

Vehicle detection based on very high-resolution (VHR) remote sensing images is beneficial in many fields such as military surveillance, traffic control, and social/economic studies. However, intricate details about the vehicle and the surrounding background provided by VHR images require sophisticated analysis based on massive data samples, though the number of reliable labeled training data is limited. In practice, data augmentation is often leveraged to solve this conflict. The traditional data augmentation strategy uses a combination of rotation, scaling, and flipping transformations, etc., and has limited capabilities in capturing the essence of feature distribution and proving data diversity. In this study, we propose a learning method named Vehicle Synthesis Generative Adversarial Networks (VS-GANs) to generate annotated vehicles from remote sensing images. The proposed framework has one generator and two discriminators, which try to synthesize realistic vehicles and learn the background context simultaneously. The method can quickly generate high-quality annotated vehicle data samples and greatly helps in the training of vehicle detectors. Experimental results show that the proposed framework can synthesize vehicles and their background images with variations and different levels of details. Compared with traditional data augmentation methods, the proposed method significantly improves the generalization capability of vehicle detectors. Finally, the contribution of VS-GANs to vehicle detection in VHR remote sensing images was proved in experiments conducted on UCAS-AOD and NWPU VHR-10 datasets using up-to-date target detection frameworks.

An algorithm for watermark based on space-domain and transform-domain is described in the paper. Making use of the different characteristic of spatial domain and transform domain, the conflict between invisibility and robust in the algorithm for watermark can be prevented, and the density of embedded watermark can be increased. In the first stage of the proposed method, the unessential parts of the watermark are embedded to the host image by LSB algorithm in spatial domain. After that, the image is converted to frequency domain with wavelet packet transform and the significant parts of the watermark are embedded to the significant coefficients. With this scheme, more watermark data can be embedded to the host image and the watermark data won't be easily distorted. From the simulation results, we can see that the approach has good performance and robust to the crop and noise attack.

In the security video surveillance system, image splicing technology faces the difficulty of achieving large parallax and small overlap area. This paper summarizes the traditional stitching technology and analyzes that the image matching algorithm is the key technology to solve the problem of large parallax stitching. Aiming at the problems that traditional matching methods with low matching accuracy cannot adapt to large parallax non-rigid transformatio, a model based on image matching is proposed in this paper for splicing tasks. The model which combines the similarity of the global structure and the local structure can not only divide the feature point on the structural level but also can pair with each other. At the same time, the framework can be trained end-to-end. The training samples composed of simulation data sets are adopted to train the designed model which is tested and verifyed by the test sample sets composed of three kinds of mixed data. Compared with PCA-GM, the model proposed in this paper adapts well to splicing tasks. The matching accuracy can achieve 93.1% and 91.9% on the simulation data set and the small parallax splicing data respectively, which can initially solve the image matching problem in the traditional image splicing scene. However, a further research aiming at the problem of the matching accuracy on the large parallax data set can only achieve 74.5% will be needed.

In this paper, a neuro-optimized numerical method is presented for approximation of HIV virus progression model in the human body. The model is composed of coupled nonlinear system of differential equations (DEs) containing healthy and infected T-Cells and HIV free virus particles. The coupled system is transformed into feedforward artificial neural network (ANN) with Mexican hat wavelet function in the hidden layers. Two meta-heuristic algorithms based on chaotic particle swarm optimization (CPSO) and its hybrid version with local search technique are exploited to tune the parameters of ANN in an unsupervised manner of error function. A comprehensive testbed is established to observe the virus growth per day with performance metric containing fitness value, computational time complexity and convergence. The proposed solutions are compared with state of art Runge–Kutta method and Legendre Wavelet Collocation Method (LWCM). The core advantages of the proposed scheme are getting the solution on continuous grid, consistent convergence, simplicity in implementation and handling strong nonlinearity effectively.

In order to ensure the security of building facades in key areas and improve the security detection efficiency of security personnel on building facades, this paper proposes an improved YOLO v3-based algorithm for open window detection recognition in building facade images, which extracts open window features from images to make predictions on full images by convolutional neural network. Firstly, due to the absence of publicly available window datasets in the network and the high number of window types that exist in reality, a self-constructed window dataset containing 13573images of open windows is used to train and test the window detection model. The data set is then clustered by the K-Means clustering algorithm to select an Anchor Box more suitable for window detection, which draws on the ShuffleNet idea to strengthen the feature extraction method, and then optimizes the network structure of YOLO v3. Finally, a block detection mechanism is introduced to effectively enhance the network’s ability to detect small dense targets; experimental results show that the method improves the accuracy and speed of window detection and reduces the workload of security personnel in key areas to manually check for windows on both sides of the street. Fire detection, missing and falling bricks on the floor, and overhead throw detectionare of great importance.

The cylinder block is the main rotating part in axial piston pump, and its dynamic characteristics can reflect the health condition of the pump more directly than the external characteristics. In this paper, a less modification measurement scheme is proposed to acquire the dynamic characteristics of the cylinder block accurately. The less modified pump minimizes the influence of the test equipment on the cylinder block motion. The radial displacements of the cylinder block surface in two cross-sections along the perpendicular directions are measured under different operating conditions. Moreover, the translational motion and the tilt behavior of the cylinder block are calculated when the run-out signals of the cylinder block surface are removed from the measurement signals. The vibration signals are measured synchronously and compared with the movement of the cylinder block. Finally, the dynamic characteristics of the cylinder block under different operating pressure are presented and analyzed in detail.

Smartphone applications became very popular nowadays as they provide useful functionalities to our daily lives over ordinary voice services. They offer a small but powerful computing platform where intelligent algorithms can be coded into some live-saving products which profoundly impact our daily lifestyles. A mobile fatigue detection system which is proposed in this study is an important life-saving application running on smartphone. Mobile detection system faces technological challenges such as: (i) embracing relatively low-resolution images in image recognition; (ii) supporting fast response time given a low-power CPU in comparison to a desktop computer; and (iii) demanding for high prediction accuracy by light-weight machine learning algorithms, as the software programs embedded in smartphones is resource constrained. Solutions with respect to indoor facial profiling system which are mainly based on progressive locating method for eye detection are discussed in this study. Acceptable experimental results in terms of eye detection rate and driver fatigue detection in different situations are presented.

The high-risk behaviour of pigs from standing, sitting to lying and grovelling is the main reason that the piglets are always crushed to death. This paper conducts research based on the field of classification and recognition of pig behaviours by using triaxial acceleration sensor, which can evaluate the maternal ability of pigs and provide data basis for selecting high quality breeding pigs. Aiming at the problem of low data variability and small data range due to the small-scale activity of the pigs, which can cause poor classification accuracy. This paper first performs moving average filter processing on the x, y, and z-axis data collected by the triaxial acceleration sensor. After feature extraction and feature selection, an optimal feature subset is proposed. Experiments show that by adopting the optimal feature subset proposed, the random forest classifier adopted can classify and evaluate four basic behaviours of pigs in daily life better. The accuracy on the test set reaches 93.8%. Compared with decision tree and BP neural network, the AUC value of random forest reaches 0.957, which has obvious performance advantages. Finally, this paper also proposed an evaluation model of maternal ability for pigs and adopted maternal ability index to evaluate the maternal ability of the pigs according to the classification results.

In order to protect the visual effect of 3D model, watermarking algorithm should have good invisibility, thus a novel frequency-domain watermarking method is proposed in this paper. Robust regions of 3D model are selected as the watermarking regions in this algorithm, and watermarking regions are transformed into two-dimensional images depending on vertices distribution. Then the two-dimensional images are transformed into wavelet transform images by using wavelet transform algorithm and the watermark is embedded in low frequency coefficients of wavelet transform images. The experimental results show that the watermarking algorithm in this paper has high invisibility, and can resist common attacks such as noise, rotation, cropping and scaling.

Gene coding has always been an important and difficult problem in the course scheduling of genetic algorithm. Starting from the definition of the set and the limited conditions of the scheduling problem, this paper expounds the coding method of the scheduling problem of genetic algorithm in detail. The gene coding method in this paper is a decimal coding with special significance. The information of each course: teaching class, class time, course name, teaching teacher, classroom type and other information are encoded as chromosomes, and the coding method is more standardized and accurate. The crossover rate and variation rate of the designed genetic algorithm will automatically change with the change of fitness, and the self-adapting crossover rate and variation rate can make the algorithm fleetly get the globing greatest result. The fitness function is improved, and the traditional genetic algorithm is compared with the improved genetic algorithm. From the final results, It shows that the improved genetic algorithm is more excellent in efficiency and performance.

In this paper, a series of electromagnetic compatibility tests were carried out on a combustible gas alarm device. In order to ensure the Electrostatic Discharge (ESD) performance of the device, some targeted solutions for electrostatic protection were proposed. Finally, the device passed the test based on analysis and rectification and met the evaluation index in GB/T 1726.3-2018 well, which is very significant to improve the quality and reliability for this kind of products.

In this study, a novel method for predicting hardness of ferromagnetic alloy based on the magnetic Barkhausen noise (MBN) is proposed. A set of new frequency features of MBN and a new hardness prediction method are proposed. The new features are derived from the first and second derivative of the auto-regressive spectrum of MBN signal. The new automatic hardness prediction method include Bag-of-Words, principal component analysis and back propagate neural network optimized by ensemble learning. The experimental results of the hardness classification show that the new features are superior to the previous features—the misclassification rate using the new features is less than 0.67%, while the misclassification rate using the previous features is about 2%. The efficiency of the new method is also proved by hardness classification experiment. Compared with the traditional time-domain method and the previous frequency domain method, the misclassification rate of the new method decreased significantly from 25% to less than 1%. In addition, the new method is highly automatic, so it is more versatile than manual algorithms. The above characteristics make the proposed new method suitable for predicting the hardness of ferromagnetic alloys in practice.

Upper-limb posture recognition is of great value to rehabilitation and assessment of stroke patients. In this paper, we propose a novel method for upper-limb posture recognition. Convolutional neural network (CNN) cascade is applied to reduce the training difficulty of the algorithm. Information of depth and color is combined to eliminate the influence of complex background and illumination variation. Kinect is used to automatically acquire a large number of upper limb posture labels. The principle of coarse-to-fine runs through the whole algorithm. The overall network architecture consists of 3 levels with a total of six CNNs. First of all, color and depth images are aligned to obtain a RGB-D quad channels image. The quad-channel image is sent to level-1 cascade network to obtain a bounding box containing the upper limb cropped from the entire body. Then, the resulting bounding box is brought into level-2 cascade network and 4 sets of rough upper-limb joints coordinates are obtained. Finally, zoom in the visual field to local area of 4 key points, 4 sets of accurate coordinate is obtained by level-3 cascade network. Experimental results show that upper-limb posture is calculated by the proposed algorithm that has strong stability to both illumination and background problems.

In the past three decades, major technologies and economic developments have been made on the telecommunication sector. HF communication has now become the main feature of the telecommunication industry. An advantage of the use of HF communication is that high frequency system is inexpensive and allows long range communication with minimal equipment requirements and easy to implement in the isolated areas as well. Frequencies ranging from 1.6 MHz to 30 MHz is called High Frequency spectrum. One of the important developments in the HF communication is Automatic Link Establishment. ALE method is widely used by the military and other organizations. In this paper we have tried to describe the implementation of ALE military standard (188-141A). This standard defines a 2G HF ALE standard. We design a communication system that allows reliable data communication over HF radio by using the Matlab environment. Our work is divided into two parts. In first part at the transmitter level, we perform the encoder to encode the data bits for error correction and detection by using the golay code. After the encoder, interleave the encoded data bits and perform the 8 array FSK modulation to transmit the data. In the second part at the receiver, we demodulated the data and then deinterleave it. After the deinterleaving we perform the decoder by using the golay decoder. At the end there is a performance analysis that comprises SNR and BER plots.

Polarimetric SAR remote sensing provides an outstanding capability of oil spill detection and classification for its advantages in distinguishing mineral oil and biogenic look-alikes. In this paper, deep learning algorithms including Stacked Auto-Encoder (SAE) and Deep Believe Network (DBN) are applied to optimize the polarimetric feature sets and reduce the feature dimension through the processes of layer-wise unsupervised pre-training. An experiment was conducted on RADARSAT-2 quad-polarimetric SAR image acquired during Norwegian oil-on-water exercise, in which verified mineral, emulsions, and biogenic slicks were provided. The results show that oil spill classification achieved by deep networks outperformed support vector machine (SVM) and traditional artificial neural network (ANN) with similar parameter settings, especially when the number of training data samples is limited.

At present, in order to get a large field of view image, image mosaic technique has been widely applied in UAV remote sensing platform. The traditional mosaic system for UAV remote sensing image takes a lot of time and man power, it is difficult to complete the image stitching automatically. In the paper, an approach for geometric correction of remote sensing image without any ground control points is presented, and the SIFT algorithm is used to extract and match feature points. Finally, the weighted average fusion method is used to smooth the image after splicing and an automatic mosaic system for UAV remote sensing video images is developed. In order to verify the system, some splicing ex-periments using UAV actual aerial photography images have been done and good results have been achieved.

The projection matrix in compressive light field acquisition has specific structural features that can be utilized for more efficient compressive sensing. In this paper, we propose a new mask design scheme for projection optimization in light field compressing with a novel circulant matrix structure. A learning circulant kernel algorithm is developed and applied to compute the projection matrix optimization from a known dictionary of the compressed light field. Experimental results show that the new approach can improve the reconstruction quality of the light field significantly compared with the commonly used I.I.D Guass random mask.

As one of the most important features of human face, the eye location plays a fundamental role in many application fields. For example, by judging the eyes state, defining whether the driver is in the state of fatigue. At present, most of the eye detection algorithms in fatigue driving are based on laboratory simulation environment. Therefore, this paper is based on the real car environment, collecting the pictures from the driver simulating the driving condition. Then, using the method of region segmentation and ellipse fitting to locate and analyze the state of eyes. All experiment pictures are collected by Android phone front camera. The algorithm runs fast and is suitable for real-time processing. At the same time, there is a good performance of eye location in the situation where there exists a certain facial orientation angle.

Keyword: Android fatigue detection, face detection, AdaBoost Abstract. Fatigue driving, a hot issue in the field of transportation research, has become a worldwide problem which always leads to traffic accident. Existing detection method always based on costly hardware platform sensors. By developing an android APP using image processing algorithm, we can lower the costs and enable convenient detection. Generally, face detection is a key step in fatigue drive detection. However, correct result can't be required by Android system detection API, especially in side face detection or strong light condition. A smart-phone based algorithm is presented in this paper to deal with side face situation and detection under strong light. Specially, the algorithm is based on a number of AdaBoost classifier that can distinguish different poses, and it exploits the color skin model with illumination compensation to increase detection quality.

In order to improve the surface properties of polyurethane (PU), samples of polyurethane have been surface modified using a newly developed active screen plasma modification (ASPM) technique. The change in the surface topography was investigated by profilometry, atomic force microscopy (AFM) and scanning electron microscopy (SEM); the chemical composition and bonding structure of the plasma modified PU surface were characterized by X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared spectroscopy (FTIR); and the wettability of the modified surface was studied by contact angle and surface energy was calculated. The results show that the ASPM technique not only can effectively alter the surface topography of polyurethane from a closed cell structure to porous open structure but also result in transformation of ester groups into new O H groups in the top surface layer. The changes of surface topography, structure and chemical composition by plasma treatment have also effectively modified the wettability of the polyurethane surface.

Based on researching of Bluetooth Protocol, a kind of Bluetooth data transferring system based on MCU-controllinghas been proposed in the paper. In the system, the MCU in which the Bluetooth HCI protocol has been embedded is used to control the Bluetooth module on UART and make the Bluetooth devices in the Bluetooth network establish connection and transfer data automatically. In the hardware design, the chip C8051F020 is selected as the host controller. Because chip C8051F020 has two UART serial interfaces, it can meet the requirement of design, simplify the circuit and enhance system anti-jamming capability. The Bluetooth module used in the system is CSR Bluetooth Core 4 with Bluetooth Standard 2.0 class 1 which has the advantages of high signal sensitivity, long distance of connection and easy to use..

A 12-lead ECG data acquisition system based on ADS1298 is presented in this paper. ADS1298 is a latest 16-channel 24-bit Analog to Digital Converter (ADC) made in TI Company, which has the characters of high-precision, low power and low noise, with the internal integration of the input multiplexers, analog low-pass filter, and digital filter, etc. Combined with the high-speed low-power chip M051 as MCU, it can achieve the standard 12-lead simultaneous ECG acquisition. ECG data can also be transmitted to an upper host PC though nRF24LE1 MCU. This system greatly compact the circuit and reduce the power cost, as well as implement the wireless transmission. It has laid the foundation for the future high-accuracy portable ECG instrument.

The use of 3D avatars is becoming more frequent with the development of computer technology and the internet. To meet users' requirements, some software or programs have allowed users to customize the avatar. However, users are only able to customize the avatar using the pre-defined accessories such as hair, clothing and so on. That is, users have limited chance to customize the avatar according to their own styles. It will be of interest to users if they are able to change the appearance of the avatar by their own design, such as creating garments for avatars themselves. This paper provides an easy solution to dressing realistic 3D avatars for non-professional users based on a sketch interface. After a user drawing a 2D garment profile around the avatar, the prototype system can generate an elaborate 3D geometric garment surface dressed on the avatar. The construction of the garment surface is constrained by key body features. And the garment shape is then optimized to remove artefacts. The proposed method can generate a uniform mesh for processing such as mesh refinement, 3D decoration and so on.

The features traditionally extracted from hysteresis loops are highly sensitive to both the variation of case depth and uncontrollable factors in repeated testing cycles, thus increasing the difficulty in predicting the tensile force applied on surface-hardened steel rods with different case depths. In this study, in order to eliminate the influence of such high sensitivity, a case depth-insensitive feature (CDIF) was proposed to characterize the tensile force, and a particle swarm optimization (PSO)-optimized neural network was used to establish the correlation between the CDIF and tensile force in order to predict the tensile force applied on steel rods with different case depths. Five classical features (including remanent magnetic induction intensities, coercive force, hysteresis loss, maximum magnetic induction, and distortion factor) and the CDIF were successfully used to characterize the tensile force. Then, the linear regression model and PSO-optimized neural network model were used in turn to establish the relationship between each feature and tensile force to predict the tensile force applied on steel rods with different case depths. The CDIF was insensitive to the variation of case depth and linearly correlated with the tensile force. Even though the CDIF is affected by the unknown and uncontrollable factors in repeated testing cycles, the PSO-optimized neural network model based on it can be used to accurately predict the tensile force applied on steel rods with different case depths with a prediction error of 0.67%.

One-dimension range profile has recently been widely used as recognition feature in radar target recognition since it can be extracted easily. However, the one-dimension range profile of a target has strong dependence on the aspect-angle of the target because it is based on the model of scattering centers. On the other hand, the different classes of target samples overlap each other in the feature space, so satisfactory recognition results cannot be obtained directly for the targets under whole posture. In order to improve the recognition performance of systems based on one-dimension range profile, a method of radar target recognition based on the multi-resolution analysis theory and neural network is presented. Firstly, one-dimension range profile features under various resolutions are obtained by using multi-resolution analysis. Then, the tree-structured cascade SOM network - a hybrid neural network on the basis of the multi-resolution analysis theory and self-organizing map (SOM) - has been used as the classifier to make multi-grade judgment and recognition gradually for the one-dimension range profile features. The experimental results show that the average correct recognition rate for three kinds of targets - a propeller aircraft, a little jet aircraft and a large jet aircraft - with 0-180^o posture variation in aspect-angle is up to 92.97%.

A high-definition digital color image watermark algorithm based on QR code and DWT is studied in this paper. The basic concept, theoretical framework and the application domain of the digital watermark algorithm is introduced. From the simulation results, this algorithm can achieve better universality, and can be applied to most of the color images. Watermarked digital works can achieved better quality, and the embedded watermark is difficult to be detected. In addition, the extracted two-dimensional code image can achieve a high quality, and can be scanned through the electronic equipment and the information in it can be acquired.

Oil spill may cause serious pollution of the marine environment. Unmanned aerial vehicles remote sensing system can be used to monitor oil spill conditions. In order to identify the oil spill area on aerial image accurately, the first step is oil spill region segmentation. The paper presents an image segmentation method of oil spill area based on fuzzy C-means Algorithm. Firstly, according to the color characteristics of the oil, the paper selects YCbCr color space as the feature space. Then, the paper uses fuzzy clustering algorithm to divide the color feature space. Finally, according to oil color model, the paper selects clustering result as the segmentation results of oil spill area. Experiment show that the proposed algorithm’s accuracy for oil region segmentation of calibration attain to 95 percent. Keywords—Oil aerial image; color model; YCbCr color space; fuzzy C-means Algorithm

With the traditional Gaussian mixture model being more sensitive to light and failing to react to changes of lighting, a variances and a mean update program under local illumination and global illumination mutations are puts forward respectively in this paper. More specifically, a divergent method to update the mean of each Gaussian distribution in the background model is proposed, following the analyses of the average grey value of current image frame and the absolute difference of the average grey value of the background model. An innovative update method to update the variance of Gaussian mixture model is also presented, based on the study of the absolute value of the pixel value and mean value. Experimental results show that the algorithm can not only detect moving targets in a relatively more complete manner, but also exhibits better adaptability and robustness for outdoor lighting mutation.

This paper presents a novel method for modeling the complexity of collaborative design systems based on their analysis and proposes a solution for reducing complexity and improving the performance of such systems. The interaction and interfacing properties among many components of a complex design system are analyzed from different viewpoints and then a complexity model for collaborative design is established accordingly. In order to simplify complexity and improve performance of collaborative design, a general solution of decomposing a whole system into sub-systems and using a unified interface mechanism between them has been proposed. This proposed solution has been tested with a case study. It has been shown that the proposed solution is meaningful and practical.

In this paper, we propose the joint use of discrete wavelet transform (DWT)-based feature extraction and probabilistic neural network (PNN) classifier to classify tissues using gene expression data. In the feature extraction module, gene expression data are firstly transformed into time-scale domain by DWT and then the reconstructed signals by using wavelet transform are reduced to a lower dimensional feature space. In the module of tissue classification, the outputs of the extractor are fed into a PNN classifier, and the class labels are given finally. Some test and comparison experiments have been made to evaluate the performance of the proposed classification scheme, using the features extracted with as well as without wavelet transform processing procedure. Correct rates of 92% and 98.7% in tumour vs. normal classification have been obtained using the proposed scheme on two well-known data sets: a colon cancer data set and a human lung carcinomas data set.

Chemical warehousing ERP system development is based on familiar chemical warehousing transit, storage, trade, business change, is effective tool to strengthen enterprise management and improve the work efficiency. This system is in Visual Studio2008 development environment, using c # programming language, ASP.NET, ADO.NET and Oracle database to jointly develop a B/S structure of ERP system. Index Terms - ASP.NET, ERP system, Oracle database, ADO.NET, B/S architecture

In this paper, in order to extract the needed information from Wireless Signal Detector, some different frequently–used methods and the related parameter estimation algorithms are compared. Besides, a new method is proposed based on signal spectrum estimation of neutral net which was used to compute signal power spectrum estimation. By simulating the practical data on MATLAB, the factor of neural network can be trained to extract the signal spectrum. The velocity of spectrum estimation is far better than some other algorithm of AR model spectrum estimation method when using proposed method, and spectral estimation quality consistent.

To keep image matching method being self-adjusted is very difficult. As an important aspect of self-adjusted problem, the angle between template and image directly decide whether the matching could be success. Particularly, the result of image matching is seldom accurate when the traditional matching approach is used. In recent years, with series of new idea created, there is a great breakthrough in field of Image matching. A new self-adjusted image matching approach based on Pyramid model and polar coordinate is discussed in this paper. Polar coordinate is used to remove the implication of rotation and Pyramid model is taken to keep the real-time of algorithm. The simulation result shows that this algorithm is more stable and faster than the traditional pixel-to-pixel algorithm. Moreover, this approach is easier to carry out and become a simple step of complex analyzing system.

Compared with the traditional communication equipment, short wave communication equipment has strong anti-destroying ability. It can greatly reduce the complexity during the data transfer process. However, by the time-variation characteristic and the influence of the multipath interference, short wave communication system has some deficiencies in transmission. Therefore, it will be of great significance if we study the problems with great efforts and find a solution. The fountain code in the short wave broadband OFDM (Orthogonal Frequency Division Multiplexing) communication system will be introduced. And an excellent concatenated code composed of the Turbo and LDPC (Low Density Parity Check Code) codes will be used to run a pre-coding stage in this paper. The simulation results show that the concatenated codes are the mirror in the realization of long codes and hold the outstanding correction capability for random errors.

An algorithm of foggy image classification is presented in this paper. First, the RGB images are converted to HSI images and next we analysis the distribution of the histograms of H, S, I plane separately, from which we extract the variance of each plane under different foggy conditions as the HSI model features. Second, the dichromatic atmospheric scattering model is introduced and based on this model we develop an algorithm for computing the angular deviation of different foggy images compared to clear day image as another feature. Finally, we use this feature set to train a multi-class SVM classifier to classify four different levels of foggy images. Experiment results show that the algorithm is more than 90% accurate. Keywords-HSI model; histogram; dichromatic atmospheric scattering model; SVM

An analog and digital modulation recognition method based on support vector machine (SVM) is proposed. A multi-class classifier is designed through the reasonable use of SVM multi-class classification method. A comparison for the performances of one-against-all (OAA), one-against-one (OAO) and binary tree (BT) with the different kernels of SVM is made. Experimental results show that the Gaussian radial basis function (GRBF) kernel has better performance than others. It can be seen from the simulation result that the proposed method is correct and efficient. The scheme can achieve 93% recognition accuracy at low levels of SNR.

In this paper, we design and present a novel ring buffer transmission system based on DM6446 and FPGA. It can improve the performance of real-time sending data from FPGA to DM6446. The ring buffer not only optimizes the efficiency and accuracy of transmission, but also verifies the correctness and reliability of the data compared with traditional ways. The proposed system realizes the radio signal transmission and analysis function.

It is obvious that the characters are illegible and the printing is rough on the package of fake cigarettes compared to genuine ones. According to the differences of comparison, we can make a two-class classifier to conduct cigarette authentication. in this paper, we firstly both extract morphological features of fake and genuine cigarettes and then use them as training sample sets to train a two-class SVM classifier. This algorithm has been tested on five brands of cigarettes and it gives results with more than 95% accuracy.

In this paper, we design software of real-time data acquisition to meet the characteristics and demands of working flow of Software Defined Radio (SDR) receiver. the software is designed in multi-thread model, which can guarantee real-time performance and efficiency compared to single thread and it can be used in Windows XP to analysis intermediate frequency signal. Besides, efficient memory allocation and management can improve stability and flexibility of the software.

in this paper, we design and present a novel real-time sleeping position monitoring system. The sensing modules composed of 3-axis accelerometers are placed on object's forehead and chest to monitor the object's position during sleeping by calculating the angles between gravity vector and its three axes. System is driven by an inexpensive and low power microcontroller. In this system, we implement a proposed novel CORDIC-based algorithm on the embedded microcontroller so that the system is capable of transferring the raw data of the accelerometer from motion domain to angular domain in-line, and the system can provide the inclination or tilt angle information by itself in real-time. The sleeping position information is integrated into polysomnography (PSG) to cooperate the study of obstructive sleep apnea (OSA) syndrome.

We proposed a bidirectional 60 GHz multiband orthogonal frequency division multiplexing (MB-OFDM) wireless communication system. In this system, MB-OFDM signals can be transmitted bidirectionally on 60 GHz band. We measured the transmission's packet error rate (PER) of the system in different distances (1m–5m) and modes (data rates of 53.3 Mbps to 480 Mbps; using various MB-OFDM time-frequency codes (TFC)) to show the system's transmission performance. Furthermore, we demonstrated the system with the form of teleconference based on Ethernet.