Applied Mechanics and Materials Vols. 385-386

Abstract: In order to analyze wide frequency band corona current characteristics of the UHV bundled conductors, a corona current measurement system is developed. The corona current signal is transmitted to the lower computer through the wireless network. And the current signal collected in this way is safe and reliable. The de-noise method with wavelet analysis is introduced and compared with some common used methods. The result of the de-noising analysis of the wide frequency band corona currents of bundled conductors in the corona cage shows that the wavelet analysis method could be used to de-noise high frequency corona current in the corona cage, and the de-noising effect is apparent.

Abstract: The conventional spectrum subtraction algorithm cannot effectively suppress the noise under highly non-stationary environment and results in the remaining music noise is often heard in the enhanced speech. In order to improve the speech enhancement performance, a novel denoising algorithm is proposed, which is based on speech endpoint detection using spectrum variance and the dynamic spectrum subtraction in Bark bands. According to human auditory characteristics, the Bark bands spectrums of the noisy speech signal are firstly calculated, and the noise power spectrum of each Bark band is then tracked and estimated by the improved minima controlled recursive averaging method. This noise estimation is adjustable frame by frame and more accurate for non-stationary environment. The experiment results showed that the proposed method can suppress the noise more efficiently than the conventional spectrum subtraction and the remaining music noise is almost eliminated.

Abstract: The traditional template matching algorithm which has a high time complexity, is susceptible to noise. This paper proposed a algorithm which based on the features of continuous orthogonal wavelet Daubechies, decomposed images to be identified to multiple layers using wavelet tranformation. In order to get matching position, we select matching template in low-frequency images. Then get matched position in higher layers after doing inverse transformation to low-frequency image. Finally, accurate position of matching template will get in original images. The algorithm not only can reduce the searching time when images are matched, but also can filter out a certain amount of noise, and so reduce the noise interference.

Abstract: In allusion to the non-stationary wideband signals, a LMS adaptive filtering algorithm based on linear canonical transform is proposed. In this method, the signal is first transformed to linear canonical transform domain. By using linear canonical transform and selecting appropriate transformation parameters, characteristics of the transformed signal appear to be stationary narrow-band in the corresponding linear canonical transform domain, and then, the transformed signal is filtered adaptively with LMS algorithm in this domain. Theoretical analysis and simulation results show that the algorithm is not only to solve the problem of extracting and filtering of nonstationary signal, and can obtain better filtering performance.

Abstract: This paper describes an architecture of ANFIS (adaptive network based fuzzy inference system), to the prediction of chaotic time series, where the goal is to minimize the prediction error. We consider the stock data as the time series. This paper focuses on how the stock data affect the prediction performance. In the experiments we changed the number of data as input of the ANFIS model, the type of membership functions and the desired goal error, thereby increasing the complexity of the training.

Abstract: In the paper the author introduces FCW_MRFI, which is a streaming data frequent item mining algorithm based on variable window. The FCW_MRFI algorithm can mine frequent item in any window of recent streaming data, whose given length is L. Meanwhile, it divides recent streaming data into several windows of variable length according to m, which is the number of the counter array. This algorithm can achieve smaller query error in recent windows, and can minimize the maximum query error in the whole recent streaming data.

Abstract: The key of ankle-brachial index (ABI) measurement is accurately measurement of upper and lower extremities systolic pressure. When the blood flow is blocked, blood flow will resume a normal beat and the corresponding cuff pressure of the first trough starting point is the systolic pressure. First it detects the pulse signal main frequency by spectral analysis of the pulse signal which based on the power spectrum estimation. Regarding pulse signal frequency as the basic standard, it achieves power spectrum estimation of blood flow signal with a sliding window. Last it achieves locating the position of systolic pressure by detecting whether the blood flow signal frequency and the pulse signal frequency are the same. Through simulation and experiments, it can achieve accurate detection of systolic pressure and automatic measurement of ankle-brachial index by application of this method.

Abstract: A time delay estimator based on the fractional bispectrum is proposed; and it is suitable for the chirp signal. The proposed time delay estimation technique can outperform the conventional time delay estimation methods associated with the bispectrum in the Fourier domain under the correlated Gaussian noises at lower SNR. Simulation results demonstrate the validity of this estimation method.

Abstract: The tilt license plate correction is an important part of the license plate recognition system. Traditional correction methods are based on one theory. It is difficult to use the advantages of different approaches. We propose some methods to help improve the tile license plate correction: a bounding box selection method based on similar height and a mutual correction method based on fitted parallel straight lines. Moreover, we use wide bounding boxes to segment touched characters. If the method based on parallel lines fails, another method, such as PCA-based one, can be used for complement. Experimental results show the proposed method outperforms others.

Abstract: The traditional moment preserving (MP) may cause owe-segmentation, over-segmentation and ignorance of image details, so this paper introduced a image threshold segmentation method based on the gradient adjustment and improved MP. Firstly, the method got a new image by sharpening the original image, then acquired the initial threshold by MP for the new image, afterwards, obtained the final threshold by the image histogram mean improving the initial threshold, finally, segmented the sharpened image. Experimental results showed that the proposed method can not only well resolve the problem of owe-segmentation and over-segmentation, get better segmentation result, but also retain a wealth of details and advantage of the original MP. The algorithm was simple, efficient and fast.