Advanced Materials Research Vols. 433-440

Abstract: This paper presents a novel approach of image interpolation based on the switching of new edge directed interpolation (NEDI) and single pass interpolation algorithm ( SPIA ) and switching is based upon the % of edges present in the blocks of the image. The switching of this interpolation algorithm is block based instead of image based or pixel based. Imperially we found that NEDI methods is better applicable for smoother images (variation among the pixels is less) while SPIA method works better on detailed images (more variation among the pixels), because of the type of pixels used in the process interpolation. So, a hybrid scheme of combining NEDI method and SPIA method is used for better prediction of HR image. The proposed algorithm produces the better results for different varieties of images in terms of both PSNR measurement and subjective visual quality with low computational complexity as compare to recently developed interpolation algorithms.

Abstract: This paper presents a new image lossless compression technique for natural images. The proposed algorithm uses the switching of existing Edge Directed Prediction algorithm and gradient Adaptive Predictor (GAP ) methods. The proposed algorithm is a switching based algorithm and the criteria of switching are based upon the adaptive threshold. We know that EDP has higher computational complexity due to the estimation of LS (Least Square ) based paramter whereas GAP has relatively lower computational complexity. So, in order to reduce the computational complexity we had made a hybrid combination of EDP and GAP Method which the proposed algorithm is a generic algorithm and it produces the best results in different varieties of images in terms of both compression ratio and computational complexity.

Abstract: Using the adaptive noise canceling technology, this paper proposes a new detecting approach to harmonics and reactive currents based on neural networks with changeable learning parameters. The structure of this neural network and the adaptive weights adjusting algorithm are presented. The contradiction of the detecting speed and the precision has been settled preferably. The proposed detecting approach can be used for detecting the harmonics and the reactive currents of active power filters. The results of the theoretical analysis and computer simulation confirm the validity of the approach.

Abstract: In the traditional design of round broach ,the impact tools of broach maybe unable to achieve the requirements for the processing precision,or even occur the situation of partial fracture due to a exaggerated partial deformation.In this article,by using Pro/E we complete the 3D solid modeling and the dimensional parameterization of the impact tools for combination round and spline broach,then import the 3D model into ANSYS,to analyze and solve the whole process of load and deformation at work.It can effectively improve the machining accuracy and the reliability of round broach, shorten the design cycle and reduce cost.

Abstract: For a slant pre-crack under mixed-mode conditions, a method is proposed in which mode Ⅰ and mode Ⅱ stress intensity factors ‹K_I›_mes and ‹K_II›_mes can be directly evaluated from the discontinuous displacement along it. The effect on fatigue pre-crack deformation behavior was discussed by comparing fatigue and annealed cracks. In general, using the biggest tangential stress criterion the direction propagating from the pre-cracks can be predicted by ‹K_I›_mes and ‹K_II›_mes calculated above. So testing for bending fatigue crack propagation under mixed-mode conditions was carried out using fatigue and annealed slant pre-cracks with slant angle β=450 defined as the angle between loading and pre-crack direction in a rectangular plate. As a result, for annealed pre-crack, the estimated fracture angle θ_estcorresponded to the measured oneθ_mes ; On the other hand, for fatigue pre-crack,θ_mes is bigger than θ_est because of the compressive residual stress around the pre-crack .

Abstract: Three-phase circuit instantaneous reactive power theory has received a successful application in detecting harmonic and reactive current timely, the ip-iq and d-q method of first proposed require the phase-locked loop and low-pass filter. The paper make corresponding improvement on the basis of the method based on amplitude integration which only can detect fundamental current , propose a method which can detect the active and reactive current, and the method based on amplitude integration can detect fundamental current of single-phase circuit directly, and simulation. The simulation results show that the method can detect the harmonic and reactive current accurately and timely, and is suitable for various circumstances that three-phase power supply distortion and asymmetry and so on, and when only detecting the harmonic current, it does not require the phase-locked loop and low-pass filter, while detecting the harmonic and reactive current, increases the phase-locked loop; when the method based on amplitude integration is directly used to detect single-phase fundamental current, the phase-locked loop and low-pass filter are not required. The paper deduces the theory of the method and researches on simulation analysis.

Abstract: In this research we propose an ensemble classification technique base on creating classification from a variety of techniques such as decision trees, support vector machines, neural networks and then choosing optimize the appropriate classifiers by genetic algorithm and also combined by a majority vote in order to increase classification accuracy. From classification accuracy test on Australian Credit, German Credit and Bankruptcy Data, we found that the proposed ensemble classification models selected by genetic algorithm yields highest performance and our algorithms are effective in building ensemble.

Abstract: Electrochemical corrosion happens when stray currents leak out of buried pipeline. This will threaten the safety of pipeline operation and operators. The distribution of stray current was studied by simulation system. The results indicate that stray current intensities have the same distribution with potential gradient Esx which parallel to the metallic pipeline. The distribution curves of stray current show regular symmetry. The stray currents increase gradually along the buried metallic pipeline. It reaches maximum at the pipeline midpoint. The potential gradient Esy distribution curves in y-direction show a hyperbolic shape. For the exits of faradic electric field, stray current intensity is maximum in the beginning and it flows into the pipeline. It flow along the pipeline and parallel to the pipeline at middle. After that, the stray currents begin to flow back to cathode and the stray current intensity reaches maximum at the end of pipeline.

Abstract: Through the in-depth study of the current motion detection and tracking technologies, combined with the practical application of intelligent video surveillance, this paper improves the existing motion detection and tracking algorithm. The improved algorithm continues the characteristics of original algorithm such as simple to implement and lower computational complexity, increases its range of application, and improves the anti-jamming capability and robustness of video tracking.

Abstract: This paper reports the basic principle and implementation process of FPGA-based multi-channel analog signal acquisition module. The emphasis of the full text is the principle of Data acquisition and the system implementation process. Multi-channel data acquisition module is composed of signal conditioning circuit、multi-channel analog multiplexers、A/D conversion and the central control system. The design uses LM324 for input signal conditioning; uses AD's 16 analog input signal chip ADG706 for channel selection; uses AD's 12-bit precision A/D converter AD7492 as the core chip of acquisition module for completing the A/D conversion; uses Xilinx's programmable logic device XC2S100 as part of the core control to achieve control of each module. Under the control of the design software, this design can take advantage of FPGA control multi-channel analog switch、A/D conversion、data compilation and uploading the data to PC to store and analyze.