Abstract:In recent years, anomaly detection plays a more and more important role in the analysis and utilization of vessel trajectory data, and has become a hot research direction in the field of navigation. The aims of detecting abnormal vessel trajectory are to study the behavioral characteristics of individuals or groups vessel and find traffic patterns and traffic characteristics hidden inside. The concept and classification of abnormal behavior of vessels are analyzed mainly from the aspects of ship position and behavior, the recent theoretical research progress in detecting abnormal vessel trajectory is summarized, the advantages and disadvantages of each method used are reviewed, and the problems and challenges in the detection of abnormal vessel trajectory are discussed finally.

Abstract:In the standard spectrum clustering algorithm, the metric based on Euclidean space cannot represent the complicate space distribution feature of some data set, which might lead to the clustering result inaccuracy. While the geometric relationship between data can be described more precise by manifold space. The special expression on curved surface is researched, the feature which is more fit for measuring the distance between data is applied, and an improved spectrum clustering analysis algorithm based on the distance metric under Graasmann manifold is proposed. The similarity between data is analyzed under manifold space. The experimental results show that the proposed algorithm can cluster data set either belonging the same or different subspace more accurately, furthermore, it can cluster data set with more complicate geometric structure under manifold space efficiently.

Abstract:A set of windscreen positioning and bracket bonding detection system based on machine vision is developed in the paper.The whole design scheme of the system is described, and the system of electric control part and robot control part and machine visionpart are introduced in detail.By taking advantage of the characteristics of the image background, the backlight and double side bar at the bottom of the light with the combination of optical control strategy is used.According to the characteristics of the product, an adaptive threshold image segmentation algorithm and an improved random Hough transform algorithm are proposed to realize the automobile windshield positioning and bracket adhesion detection.The experimental results show that the designed system can well reduce the error rate of artificial bonded bracket, and the defective rate is only 4%,it meets the production precision and time requirements. The bonding precision is less than 0.5 mm, intelligent manufacturing system production efficiency is 4.5 times of artificial, the production efficiency of the enterprise is greatly improved.

Abstract:Fearful emotion is a response to the external stimuli of human. The generation of fearful emotion could lead to the change in the facial skin temperature. According to the principle of infrared thermal images reflecting the temperature distribution on the surface of objects, a method based on infrared thermal images is proposed to recognize the fearful emotion. Firstly, a heat transfer model is simplified by curving fitting of exponential function, and the facial infrared thermal image is converted into the blood perfusion pseudocolor image to find the regions of interest (forehead region). Then, features of the blood perfusion change curves are extracted (slope, confidence coefficient, mean value, and standard deviation), and the correlation between the features and the selfassessment score of the fearful emotion is analyzed using Spearman correlation coefficient. Finally, the standard deviation which is highly related to the selfassessment score is applied to recognize the fearful emotion of the subject. The experimental results show that there is an obvious decrease in the blood perfusion of forehead region in the presence of fearful emotion, which is consistent with observations of previous studies, and the standard deviation (with a threshold of 0.14) of the blood perfusion values is a main feature for recognition of the fearful emotion. The proposed method is demonstrated to be satisfactory and reliable with an accuracy of 85.7% for all the 28 tested subjects.

Abstract:In order to take full advantage of NetworkonChip parallelism ability and improve network bandwidth utilization, a highperformance dualport network is introduced in this paper. Each router in the structure provides two local ports, and each resource node connects with two routers in a diagonal. A destination node switching method is proposed in this structure to improve the parallel communication ability of network. Meanwhile, to further improve the performance of the network, the Torus structure is introduced to reduce the network radius. This paper builds a model of dualport network, and evaluates its performance. Compared with signalport network, the experiment results show that the structure can improve the average throughput and the average latency by 83.3% and 559% at best in unary affairs. Compared with the same dimension dualport network, the experiment results show that the structure can improve the average throughput and the average latency by 91.1% and 54.3% at best in binary affairs.

Abstract:In order to meet the double requirements which are the largescalein measurement range and highprecision in local detail in the manufacturing and assembly process of marine engineering,a method to construct the global control network and optimize the transfer station is presented in this paper.Based on the research on the rules of the measurement process of laser tracker, the mathematical model of measurement process is established, and the uncertainty of single point measurementis predicted according to the result.The experiments show that the actual measurement and simulation data are in good agreement. Therefore, the point measured by the tracker different weight distribution is given, andwhich is integrated into the SVD decomposition method,then the parameters of transfer station can be obtained. And compared with the ordinary SVD in RMS, the SVD method with the weight has a statistically significant optimization.Transfer station parameter is used to analyze the transfer station accuracy.The measurement errors existed in bothtwo stations are considered with TLS (total least squares),the affecting factors of precision are figured out, it laysthe foundation for the followup study to improve the precision of the transfer station.

Abstract:Aiming at the inspection of genetically modified product is mainly based on visible or near infrared spectroscopy at present, the support vector machine (SVM) modeling parameter is difficult to determine and the problem of the large amount of spectrum data calculation, a support vector machine (SVM) algorithm based on terahertz spectroscopy and adaptive particle swarm optimization (APSO) is proposed to distinguish genetically modified cotton seed. To achieve distinguish genetically modified cotton seed, the present invention is train of thought collect 165 samples of three kinds of latest genetically modified cotton seed of terahertz spectroscopy in range of 150 μm~3 mm wavelength and identification of 165 genetically modified cotton seeds based on APSOSVM. The experiment results show that the comprehensive recognition rate reached 97.3%. It can provide a precise, fast, convenient, and nondestructive detection method to distinguish genetically modified cotton seed by using terahertz spectroscopy couple to APSOSVM.

Abstract:A denoising scheme based on wavelet shrinkage technique is proposed to remove the noise in the heart sound signal. Firstly, the characteristics of heart sound signal frequency were analyzed and Haar, Daubechies, Symlets and Coiflets orthogonal wavelets were studied for contrast in accordance with the principle of frequency band similarity matching. Based on the statistical results, Coif5 wavelet was chosen for the decomposition and reconstruction of heart sound signal. Besides, a smooth and continuous adaptive elastic threshold function was designed for wavelet shrinkage, which can overcome the problem of discontinuous hard threshold function. The noise reduction effects were compared under four threshold rules. The simulation results show that, when the SNR is less than 50 dB, the optimization scheme with Heursure threshold rule can retain sufficient heart sound detail information, while effectively removing noise.

Abstract:In this paper, a tracking algorithm based on CAMShift which combined with SURF feature matching and Kalman filter is proposed to deal with the problems in traditional CAMShift algorithm, such as tracking failure under color interference or occlusion. The algorithm calculates the Bhattacharyya coefficient of integrated histogram composed of chroma feature and gradient direction feature between candidate target and template target as judging basis, it uses CAMShift algorithm. As the coefficient more than the threshold, SURF algorithm will be used to match the search window and the tracking result of the previous frame, then recalculate the target’s size and position by the matching result. To avoid tracking failure by fastmoving of target and reduce the computation of SURF matching, the center position of moving target in the next frame will be predicted by Kalman predictor. The experimental results show that the new algorithm can achieve stable tracking object against complex backgrounds, color interference or occlusion, and have higher tracking speed than CAMShift algorithm combined with SURF.

Abstract:A new method to evaluate the freeform surface error is proposed in this paper. Testing whether the error is within the required tolerance zone is implemented by comparing the reconstructed machined surface with the design surface. The inspection process is created out by these steps: acquisition of measurement points on the actual surface, transform the measurement points from the measurement coordinate system to design coordinate system by location; the machined surface is reconstructed by the located points based on Bspline interpolation; the reconstructed surface is compared with the design surface. Because of the uncertainty in measurement process, the uncertainty of location, reconstruction, and form error are analyzed. The proposed method is verified in the measuring and analysis of “S” shape test shape. The results show that the proposed method is effective to guide the evaluation and acceptance of free form surface machining quality.

Abstract:Aiming at the problem that most conventional methods cannot achieve high precision when ionospheric delay variation increasing at low sampling rate, a new cycleslip detection and correction method is proposed. Firstly, a geometryfree and ionospherefree carrier phase combination (GIF), and a secondorder timedifference phase ionospheric residual combination (STPIR) are integrated to detect the cycleslips epochs. Then in these cycleslips epochs, two ionospherefree codephase combinations are selected as the first and the second detectable amount respectively, and a geometryfree combination (GF), which ionospheric variation is taken into account, is selected as the third. These three combinations are composed linearly independently to find the values of cycleslips. Finally, the experiment with BeiDou triplefrequency is given to verify the method. The experiment indicates that even in 30 s sampling intervals, the method can still detect different types of cycleslip, especially insensitive cycle slips. This method can effectively reduce the effect of increasing ionospheric delay variation at low sampling rate, and be applied to dynamic, undifferenced observations.

Abstract:Aiming at the problem of testing the ABS gear ring round surface defectsby the traditional manual testingmethodswithlow efficiency, easy to error testing and missed testing,a method for testing the ABS gear ring round surface defect based on image processingisproposed.According to the actual produce, an on line visual testing system is designed and assembled. Using electric rotating machinery and gear ring supporting platform to drive gear ring around and get the picture of the gear ring round surface by liner CCD scanning. After processing the picture by image algorithm based on OpenCV,thedefectsarejudgedaccording to different existingarea,andthen the qualification of the gear ring is judged. Comparing with the results of system testing and artificial testing, itshows that the average testing time of each gear ring less than 4 s, the success rate of defect classification is more than 92%.

Abstract:For vibration signal generated by rubbing fault in rotary machinery including the singlepoint rubbing and the partial rubbing, multicomponent harmonic frequencies are presented. An ensemble empirical mode decomposition (EEMD) method was introduced to resolve the difficulty of multicomponent fault feature extraction of rubbing fault. The rubbing fault signals created by simulation were decomposed clearly into different single frequency data by using EEMD method. The vibration signals from experiment rig of rubimpact rotor were then analyzed under different directions rubbing, which include in horizontal, in vertical, in horizontal and vertical at the same time. The vibration strength belonged to different fault frequency in the frequency spectrum was obtained from the intrinsic mode function (IMF) based on EEMD respectively. The analysis results of the rubbing signals indicate that the multiple features can be better extracted with the EEMD, and is effective in analysis and recognition of rubbing fault.

Abstract:In the paper, a polyethylene terephthalate (PET) bottle tamperevident band broken detection algorithm based on contour curvature and corner detectionis proposed to solve the problem that no existing algorithm concentrates on PET bottle tamperevident band breakage detection.Firstly, the bottle cap area is roughly located by grayscale projection and the rectangular area ofthe cap is set to region of interest(ROI). Then, the cap edge point is obtained by gradient maxima suppression method,the subpixel edge is fitaccording to the given law isosceles triangle, and the bottle cap contour is connectedthrough the nearest point search method.And then,the stretch to distance ratio(SDR) of edge point is used to approximate the contour curvature,the corner points are detectedbased on maximum contour curvature, and the support ringis locatedby corner matching method. Finally, whether the tamperevident band broken is determined according to whether or nota gap exist between supporting ring and the tamperevident band.The algorithm detection correct rate is 94.75%,it meetsproduction needs.

Abstract:Aiming at the problem of parametric fault diagnosis in nonlinear circuits, an approach utilizing cepstrum and decision tree is proposed. Firstly, the acquired fault response signals are converted by cepstrum. Then, the wavelet analysis is used to decompose the converted data and the energy is taken from different frequency bands. Finally, the obtained fault features are inputted into decision tree to identify different faults. The simulation results show that the proposed method can extract the fault signature effectively and can get a good diagnosis result.

Abstract:Some edges cannot be detected when the single property gradient edge detection is used,it affects the integrity of the detection results. To solve the problem, a new edge detection algorithm based on multigradient fusion is proposed. The algorithm first converts the RGB image into the YCbCr color space, and then extracts the luminance component Y and the color component Cb, Cr respectively. For the extracted luminance component Y, the color components Cb and Cr, combined with the depth image D, using the multidirectional circular edge detection operator calculates the gradient respectively.The four gradient images are fused to obtain the edge detection results. The experimental results show that the algorithm gains clearer and more complete results, and improves the effect of edge detection effectively.

Abstract:When counting papers and wafer cells, the sides of gathered images are horizontally striped. During its filtering, traditional filtering algorithm often results edge image blur. To solve those problems, this paper presents an adaptive template filtering algorithm for vertical solar cells image, which determines threshold by the image gray, finds edge pixels to exclude, and creates the template for the rest of the pixels by surrounding pixels’ gray value. It is seen from the graph of the gradation values before and after filtering that the algorithm has good effect at the edges, and the gray of result image is highly smoothed, and the algorithm has practical value.

Abstract:To research the audio classification of “memory of the world heritage”Dongba classic books, the way of speech emotion feature extraction is adopted to identify the Dongba audio categories and realize the emotional state recognition of Dongba audio.And to improve the performance of human computer interaction, the way of speech emotion feature extraction based on Mel frequency cepstrum coefficient(MFCC) is adopted.The first order difference and the second order difference are introduced to describe the dynamic characteristics of speech features. The speech signal characteristic parameters based on MFCC and the shorttime energy are finally formed to extract the characteristics of speech emotion feature. The experiment research shows that the characteristics of the speech signal can be more clearly differentiate emotional information contained in the speech, and lay the foundation for recognizing speech emotion and the Dongba audio classification.

Abstract:In this paper, a traffic sign recognition algorithm based on logpolar transformation and Zernike moment was presented. First, to improve the image contrast, histogram equalization was performed in HSI color space towards the image captured from complex natural environment. After that, traffic sign was detected by color, and segmentation and region merging was carried out.Next, it’s followed by screening by shape and subsequent normalization. Then, images’ Zernike moment was computed combining logpolar transformation. Lastly, SVM classifier was used to recognize the object. The experiment result shows 94.71% detection accuracy and 85% recognition accuracy, which demonstrates that the traffic sign recognition system can effectively recognize the distortional, scaling or rotated traffic signs.

Abstract:Thermal analysis instruments usually have high performance requirements, and the study of the temperature field distribution of the furnace is the prerequisite for the realization of precise econometric analysis.Due to the influence of air flow on the temperature distribution, the traditional furnace is prone to measurement error. Adjustingthe mechanical structure of furnace can improve the accuracy of the equipmentand change its temperature drift characteristics.In this paper, a threedimensional optimal distribution model of furnace body is established.And the finite element analysis software is used to analyze the temperature field.In this paper, with the analysis of thermodynamics, heat transfer methods, temperature field distribution under different inlet flow, the relationship between the temperature field distribution of the furnace and the gas flow, and the direction of the furnace is described.It provides a theoretical basis for improving the accuracy of temperature control of the thermal analysis instrument.The analysis method and theoretical results can be used as a reference for distribution of thermal analysis instrument furnace temperature field and uniform temperature region in the furnace. And then it provides the basis for the research of the temperature control of the thermal analyzer. And it can lay a foundation for improving the repeat ability, accuracy and stability of the thermal analysis instrument.

Abstract:High intensity focused ultrasound (HIFU) can irradiate fresh pork in vitro, which causes 3 degrees of lesion of pork tissue. From the aspects of Bmode ultrasound image processing, the research on biological tissue lesion level judgment based on Kmeans clustering and combined with double parameters is proposed in this paper. Realtime Bmode ultrasound images of 134 pork tissues before and after HIFU irradiation can be obtained by B ultrasonic instrument, and they are preprocessed to get digital subtraction images of the focal spot area. Then the gray average and the mean of the wavelet transform coefficient of these digital subtraction images can be extracted. Meanwhile, the pork tissue samples can be classified by Kmeans clustering. The results show that gray average can distinguish the second and the third level of tissue lesion more effectively, and the mean of the wavelet transform coefficient can distinguish the first and the second level of tissue lesion more effectively. However, the method based on Kmeans clustering and combined with double parameters is equipped with the advantages of the two former. And compared the two formers, this method improves the recognition rate of tissue lesion level by 5.23% and 3.43% respectively. And it can judge the lesion level of the pork tissue more accurately. The method can help clinicians to monitor the HIFU treatment process objectively, and it has practical significance to improve the HIFU therapeutic effect.

Abstract:An UAV Attitude Control System based on least squares fit and quaternion expand Kalman filter estimation scheme is proposed. Firstly, the physical model of triaxial magnetometer is established, the least squares fitting algorithm is used to estimate the disturbance vector magnetometer, and then the accelerometer is used to tilt compensation magnetometer. Finally, the 3D pose of the aircraft is estimated by expanding the Kalman filter fusion accelerometer, gyroscope and magnetometer data. The test results show that the maximum linearity error of the azimuth is 4 °, tilt angle is 40 °, the maximum azimuth error is 2.6°, and in static placement, the maximum static deviation of roll angle, pitch angle and azimuth is 0.215 °, 0.103 ° and 0.464 ° respectively.

Abstract:In order to resolve the differential impedance of electric connector is relatively small and discontinuous at the R & D stage, the optimization scheme which is composed of the replacement of the dielectric material and the modification of the contact and the medium size based on the theory of signal integrity are proposed. The HFSS high frequency electromagnetic field simulation software was used to optimize the contact pin diameter（W1）and fit lengths of contact and dielectric (L1). When the pin diameter is 1.1 mm, the simulation value of the differential impedance of connector is 100±10 Ω and L1 has little influence on the simulation value of differential impedance of connector. The improved connector was processed and the differential impedance test was carried out. The test results show that the differential impedance range of connector is from 84.46 to 90.68 Ω before improvement and from 92.6 to 105.3 Ω after improvement, differential transmission eye diagram meets typical highspeed transmission electric connector signal integrity transmission requirements after improvement.

Abstract:The quartz crystalunittesting method of zero phase technique in a πnetworkrecommended by IEC relies on strict limits on test environment to ensure the test accuracy. In practical application, the test environment of production site will bring in stray reactance, which seriously affects the test accuracy. To improve test accuracy, according to the model of electrical parameter of quartz crystal as well as the difference between idealπnetwork and actualπnetwork model, the effect of distributed inductance on the test accuracy of resonance frequency of quartz crystal were explored based on theoretical analysis and experiments. The method of capacitance compensation could significantly reduce the effect. The experimental results show that the testing accuracy of series resonance frequency of quartz crystal could be up to ±2×10-6after compensation.