Abstract: Within the framework of computational auditory scene analysis (CASA), a speech separation algorithm based on energy difference for close-talk system was proposed. The two microphones received the mixture signal of close target speech and far noise sound at the same time. The inter-microphone intensity differences (IMID) of the two microphones in time-frequency (T-F) units were calculated. And used as cues to generate the binary masks with the K-means two class clustering method. Experiments indicated that this novel algorithm could separate the target speech from the mixture sound, and performed well in a big noise environment.
Abstract: This article introduces a vehicle license plate localization method which is based on gray image. This method mainly processes some uneven illumination and low-quality vehicle images. Firstly it uses edge detection operators to enhance images to plate region. Then it enhances with algorithm such as binaryzation, first-order points difference and median filtering in custom templates. And lastly it makes it more accurately in license plate localization through projection method enhancement. It is showed in experiments that this method can license plate localizes these kinds of low-quality license plate images accurately and effectively.
Abstract: Transformation used for aligning feature-based scans pair, is pivotal in optimization and simplification for splat-based scans. In this paper, a novel optimal transformation algorithm is first presented by take advantage of transformation between two coordinate system which is directly determined by position and normal. And then transformation is simplified by eliminating the redundant transformation to remit the computation complexity. At last, we give the transformation result of apple scans registration in our experiment platform and verify the accuracy and performance in fast computation of algorithm.
Abstract: The scene distance information is obtained by disparity based on binocular stereo vision. It is widely used in scientific research and practical engineering. Two cameras which fixed on ahead of the vehicle are used to capture stereo image pairs. Then the backwindows of the two images are located using Hough transform. The top left corner of the windows can be treated as a pair of matching points to calculate the distance of the front vehicle according to the principle of binocular stereo vision. The distance measurement scheme is presented. The experimental results demonstrated that the algorithm provides accurate localization and can help driver to deal with traffic problem to some extent.
Abstract: To overcome the problems of traditional PWM inverter, the paper puts forward a novel parallel resonant inverter, and expounds the inverter’s configuration and working principle. It analyses choice tenet of resonant inductance and resonant capacitance of resonant link. It confirms the parameters through calculation accordimg to the need of system. It simulates the resonant link of different parameters, comparison of the simulation results testifys validity of the resonant inverter.
Abstract: On the basis of the analysis of spatial distribution characteristics of massive ship tracks data, the model of spatiotemporal cube was established and the “3DR-Tree” algorithm was proposed to study the efficient index building technology. The spatiotemporal query of massive ship tracks data was realized and the ship historical trajectories were also displayed in the global environment intuitively and dynamically, it can be used for maritime search and rescue effectively.
Abstract: Focusing on the nonlinear and uncertain characteristics of suspension system,a 2-DOF vehicle is regarded as the control object, sliding mode theory was used to design a sliding mode controller for the 2 DOFs vehicle semi-active suspension system,then RBF neural network was employed to optimize the sliding mode controller.The control effects of three key performance parameters of suspension, the acceleration of car body, the dynamic travel of suspension and the dynamic deflection of tire are studied under random excitation conditions.The results indicate that in comparison with the passive suspension，sliding mode semi-active control based on RBF neural network can improve suspension performance effectively.
Abstract: For the single-channel autoregressive moving average (ARMA) signals with multisensor, and with unknown model parameters and noise variances, the local estimators of unknown model parameters and noise variances are obtained by the recursive instrumental variable (RIV) algorithm and correlation method, and the fused estimators are obtained by taking the average of the local estimators. Substituting them into the optimal fusion Kalman filter, a self-tuning fusion Kalman filter for single-channel ARMA signals is presented. A simulation example shows its effectiveness.
Abstract: Noise mixed in the recorded seismic signals often affects the data analysis result. To solve this problem, wavelet packet decomposition technology is selected to reduce the noise. The principle, denoising steps and algorithm of wavelet packet decomposition technology are introduced with its effectiveness and reliability verified by computer simulation. Meanwhile, split-radix FFT algorithm is used to rapidly perform spectral analysis for seismic signals. Compared with radix-2 FFT and radix-4 FFT, split-radix algorithm can significantly reduce calculation amount and improve the speed of data analysis. Field tests showed that wavelet packet denoising technology and split-radix FFT algorithm have good effect.
Abstract: As an important element of the index in room acoustics, the reverberation time has been considered as an objective parameter with clear concept and had good correlation with the subjective perceptions in the field of psycho-acoustic criteria. Through the computer acoustic simulation software EASE and the electro-acoustic test system CLIO, this paper conducts research on the reverberation time of the car speaker sound field, which in the domestic car of a certain type, studies show that EASE is accurate and reliable in the simulation of the car speaker sound field, the reverberation time is greater than 0.4s when the frequency less than 500 HZ and less than 0.1s when the frequency greater than 1000 HZ, when reaches to 500 HZ, the reverberation time reaches the maximum 1.103s.