Applied Mechanics and Materials Vols. 536-537

Abstract: In this paper, aiming at the poor performance of the existing symbol rate estimation methods when SNR is low, we proposed a method based on the timing square which can be simply calculated and applied to MASK/MPSK/MQAM modulated signal, improved the performance of signal symbol rate estimation compared with the existing method. Based on the principle of the timing square, the signal modules were squared to calculate the fourier coefficient modulus corresponding to different sampling ratios. Then the characteristic line was searched in the transform spectrum which contains symbol rate information and obtained an estimation of the symbol rate. The impact of the roll-off factor and the carrier wave was analyzed and solution was proposed in this paper. Simulation results show that the performance of symbol rate estimation of the improved method is better than the original method, the wavelet transform method and the cyclostationary method in low SNR and low roll-off factor environment.

Abstract: Fault diagnosis is essentially a kind of pattern recognition. In this paper propose a novel machinery fault diagnosis method based on supervised locally linear embedding is proposed first. The approach first performs the recently proposed manifold learning algorithm locally linear embedding on the high-dimensional fault signal samples to learn the intrinsic embedded multiple manifold features corresponding to different fault modes. Supervised locally linear embedding not only can map them into a low-dimensional embedded space to achieve fault feature extraction, but also can deal with new fault samples. Finally fault classification is carried out in the embedded manifold space. The ball bearing fault signals are used to validate the proposed fault diagnosis method. The results indicate that the proposed approach obviously improves the fault classification performance and outperforms the other traditional approaches.

Abstract: This paper proposes a novel fractional-order Laplacian operator for image edge detection. The proposed operator can be seen as generalization of the second-order Laplacian operator. The goal is to utilize the global characteristic of the fractional derivative for extracting more edge details. A thresholding is set based on the average fractional-order gradient for marking the edge points, and then the image edge can be extracted. Experiments show that the proposed fractional-order operator yields good visual effects.

Abstract: A new color fidelity algorithm to reduce sensor noise for high-ISO images has been proposed. The new algorithm uses HSV color space on wavelet domain to denoise high-ISO images by a weighted mean filter, which is designed based on analysis of the different noise models of high-ISO images. Testing results are satisfactory, which have shown that the new algorithm could reduce noise in high-ISO images faster, as well as no any color distortion in original images during the denoising processing.

Abstract: Video denoising is an important task to enhance video quality in many applications. The recently proposed Polyview Fusion (PVF) is one of the best methods that significantly improves results by applying 2-D denoising algorithm to multiple views (front, top and side views) and fusing the denoised results into one. In general, denoised result in smooth content is better than that in texture content; and if there is no significant motion in the video signal, the top or side view is mainly consisted of smooth content. Based on it, a new fusion method is proposed. Three denoised results are transformed into top or side views. And then they are divided into different regions and given different weights based on local features. After fusion and transformation, it obtains the final front view result. Experimental results confirm the better performance of the proposed method for the video sequences without significant amount of rapid motion.

Abstract: This paper proposes a FPGA implementation to apply a stereo matching algorithm based on a kind of sparse census transform in a FPGA chip which can provide a high-definition dense disparity map in real-time. The parallel stereo matching algorithm core involves census transform, cost calculation and cost aggregation modules. The circuits of the algorithm core are modeled by the Matlab/Simulink-based tool box: DSP Builder. The system can process many different sizes of stereo pair images through a configuration interface. The maximum horizon resolution of stereo images is 2048.

Abstract: In this paper, we deal with robust detection of defects that can occur on fastener heads peripheral side. In machine vision, the first necessary step is design good illumination system that can give image with good contrast between good and bad samples. In our case, it is difficult to devise such a god illumination system. First, we find the outer boundary of fastener head using binarization, connected component analysis and checking local distribution of intensity. Then four control points on outer boundary is chosen to ease finding the inner boundary of fastener head using polygon approximation and line fitting. Finally, decision can be done using two detected boundary points of fastener head.

Abstract: Video compression coding technology is the key to implement network video transmission, and the increasing requirements of network content transmission quality accelerate the development of video compression coding technology. H.264 is a highly compressed digital video codec standard, and is a hybrid coding mode of differential pulse code modulation (DPCM) and transform encoding. In this paper, we introduced several techniques of hybrid coding algorithm based on H.264 standard, including differential coding, block matching motion prediction method, normalization criteria, variable-size block motion compensation, and multiple hypothesis motion compensation. We did extensive experiments on variable-size block motion compensation coding method. The experimental results show that, with the premise of the 1/4 pixel motion compensation accuracy and hybrid coding, 8 × 8 or larger block partition of content for normal video object, can further make use of space correlation and improve compression efficiency.

Abstract: Shape recognition methods based on one-dimensional global feature Fourier descriptors have been widely used for the computational simplicity and high recognition rate. However, in cases of occlusion, local deformation or other complex degradation this approach shows poor performance. In this paper we present a new shape recognition method with a novel shape similarity metric, which fuses three different metrics, all of which are rotation, translation and scaling invariant. These metrics takes into account the normalized shape variance, phase correlation and modified Fourier descriptor. Experimental results show that the proposed method has high recognition rate even when the object undergoes some shape deformation. When compared with other traditional shape recognition methods, the proposed method demonstrates significantly improved performance.

Abstract: In this paper for LBG algorithm is very sensitive to initial code book, clonal selection algorithm is introduced and splitting method is adopted to generate the initial code book. Besides, because similarity metric based on euclidean distance can only reflect local consistency of the clustering instead of the clustering global consistency, clonal selection clustering method in terms of manifold distance is put forward to produce and optimize the final code book. Furthermore, it is proved through the experiment that algorithm of clonal selection code book design based on themanifold distance(MDCSA) is characterized by better performance .