De Xiang Zhang

Video Tracking of Human with Occlusion Based on MeanShift and Kalman Filter

In order to accomplish tracking of moving objects requirements, and overcome the defect of occlusion in the process of tracking moving object, this paper presents a method which uses a combination of MeanShift and Kalman filter algorithm. MeanShift object tracking algorithm uses a histogram to describe the color characteristics of an object, and search the location of an image region that the color histogram is closest to the histogram of the object. Histogram similarity is defined in terms of the Bhattacharya coefficient. When the moving object is a large area blocked, the future state of moving object is estimated by Kalman filter. Experimental results verify that the proposed algorithm achieves efficient tracking of moving objects under the confusing situations.

Fusion of Visual and Infrared Image Based on NSCT

Non-subsampled contourlet transform (NSCT) is the combination of the multi-scale analysis and multi-directional analysis in processing high-dimensional signals and has better approximation precision and better sparse description. A novel and efficient fusion method of infrared image and visible image based on NSCT is proposed. Firstly, source fusion images can be decomposed into low-frequency coefficients and high-frequency coefficients using the NSCT. For the low-frequency coefficients, the average fusion algorithm is used. For the each directional high frequency sub-band coefficients, the larger value and region average gradient maximum criterion is used to select the better coefficients for fusion. Experimental results show that the proposed algorithm can achieve better result compared with the traditional image fusion algorithms.

Research on Polarization Image Fusion Using Tetrolet Transform

Tetrolets are Haar-type wavelets whose supports are tetrominoes which are shapes made by connecting four equal-sized squares. Firstly, the magnitude of polarization image and angle of polarization image can be decomposed into low-frequency coefficients and high-frequency coefficients with multi-scales and multi-directions using tetrolet transform. For the low-frequency coefficients, the average fusion method is used. For the each directional high frequency sub-band coefficients, the larger value of region energy information measurement is used to select the better coefficients for fusion. At last the fused image can be obtained by utilizing inverse transform for fused tetrolet coefficients. Experimental results show that the proposed method can detect image features more effectively and the fused image has better subjective visual effect.

Fusion of Polarization Image Based on Contourlet Transform

The contourlet transform is a new two-dimensional extension of the wavelet transform using multiscale and directional filter banks and has better approximation precision and better sparse description. Firstly, the magnitude of polarization image and angle of polarization image can be decomposed into low-frequency coefficients and high-frequency coefficients with multi-scales and multi-directions using the contourlet transform. For the low-frequency coefficients, the average fusion method is used. For the each directional high frequency sub-band coefficients, the larger value of region teager energy information measurement is used to select the better coefficients for fusion. At last the fused image can be obtained by utilizing inverse transform for fused contourlet coefficients. Experimental results show that the proposed algorithm works better in preserving the edges and texture information compared with the traditional image fusion algorithms.

Fusion of SAR Image Using Stationary Contourlet Transform

The stationary contourlet transform is built upon nonsubsampled pyramids and nonsubsampled directional filter banks and provides a shift invariant directional multiresolution image representation. Firstly, several SAR images can be decomposed into low-frequency coefficients and high-frequency coefficients with multi-scales and multi-directions using the stationary contourlet transform. For the low-frequency coefficients, the average fusion method is used. For the each directional high frequency sub-band coefficients, the larger value of horizontal and vertical direction gradient information measurement is used to select the better coefficients for fusion. At last the fused image can be obtained by utilizing inverse transform for fused contourlet coefficients. Experimental results show that the proposed algorithm gives more satisfactory results than the traditional image fusion algorithms in preserving the edges and texture information.

Wavelet Denoising of SAR Images Corrupted by Speckle Noise Using Cycle Spinning

A novel and efficient speckle noise reduction algorithm based on wavelet transform by cycle spinning for removing speckle of unknown variance and minimizing the effect of pseudo-Gibbs phenomena from Synthetic Aperture Radar (SAR) images is proposed. Therefore, we show that the sub-band decompositions of logarithmically transformed SAR images. Then, we process and reconstruct multi-resolution wavelet coefficients by wavelet-threshold using cycle spinning, a technique estimating the true images as the linear average of individual estimates derived from wavelet thresholded translated versions of the noise images. Experimental results show that the proposed de-noising algorithm is possible to achieve an excellent balance between suppresses speckle effectively and weaken as many image Gibbs phenomena as possible. Quantitative and qualitative comparisons of the results obtained by the new method with the results achieved from the other speckle noise reduction techniques demonstrate its higher performance for speckle reduction in SAR images.

Speech Stream Detection for Noisy Environments Based on Empirical Mode Decomposition

A new approach for speech stream detection based on empirical mode decomposition (EMD) under a noisy environment is proposed. Accurate speech stream detection proves to significantly improve speech recognition performance under noise. The proposed algorithm relies on the Teager energy and spectral entropy characteristics of the signal to determine whether an input frame is speech or non-speech. Firstly, the noise signals can be decomposed into different numbers of sub-signals called intrinsic mode functions (IMFs) with the EMD. Then, spectral entropy is used to extract the desired feature for noisy IMF components and Teager energy is used to non-noisy IMF components. Finally, in order to show the effectiveness of the proposed method, we present examples showing that the new measure is more effective than traditional measures. The experiments show that the proposed algorithm can suppress different noise types with different SNR.

Speech Endpoint Detection Based on EMD and Spectral Entropy in Noisy Environments

Accurate endpoint detection is crucial for speech recognition accuracy. A novel approach that finds robust features for endpoint detection based on the empirical mode decomposition (EMD) algorithm and spectral entropy in a noisy environment is proposed. With the EMD, the noise signals can be decomposed into different numbers of sub-signals called intrinsic mode functions (IMFs), which is a zero-mean AM-FM component. Then spectral entropy can be used to extract the desired feature for IMF components. In order to show the effectiveness of the proposed method, we present examples showing that the new measure is more effective than traditional measures. The experiments show that the proposed algorithm can suppress different noise types with different SNR, and the algorithm is robust in the real signal tests.

Gearbox Fault Diagnosis Based on EMD and Coefficient-Energy Value

This paper applies the empirical mode decomposition (EMD) methods to gearbox vibration signal analysis capture from vibrating acceleration sensor for gearbox fault diagnosis. The original modulation fault vibration signals are firstly decomposed into a number of intrinsic mode function (IMF) by the EMD method. Then the fault information diagnosis of the gearbox vibration signals can be extracted from the coefficient-energy value of intrinsic mode function. Experiment result has shown the feasibility and efficiency of the EMD algorithms and energy characteristic method in fault diagnosis and fault message abstraction. It is significant for the monitor operating state of gearbox and detects incipient faults as soon as possible.

Gearbox Fault Diagnosis Based on Empirical Mode Decomposition and Hilbert Transform

Gearbox is vital components in a wide range of industrial and transport applications. It is very important how to monitor operating state of automobile gearbox and detect incipient faults. This paper applies the empirical mode decomposition (EMD) and Hilbert spectrum methods to gearbox vibration signal analysis capture from vibrating acceleration sensor for gearbox fault diagnosis. The original modulation fault vibration signals are firstly decomposed into a number of intrinsic mode function (IMF) by the EMD method. Then Hilbert spectrum of intrinsic mode function at different fault characteristic frequencies is obtained by Hilbert transform. Finally, the time-frequency fault characteristics of gearbox are analyzed by the Hilbert spectrum value of intrinsic mode function. Experiment result has shown the feasibility and efficiency of the EMD algorithms and Hilbert spectrum characteristic method in fault diagnosis and fault message abstraction.