Applied Mechanics and Materials Vol. 742

Abstract: The process of Pichia pastoris fermentation has a long period and less offline data . The cell concentration and some other important variables can not be measured on line.The soft sensor modeling at present is mainly the artificial neural network (ANN).This paper introduces the Support Vector Machine (SVM). Select fewer off line data and establish soft sensor modeling about cell concentration.Compared with the BP neural network prediction model, Simulation experiment proves that the support vector machine has better prediction effect and generalization ability. template.

Abstract: Combined with the characteristics of forest image, the paper puts forward a new forest fire image segmentation method based on YCBCR and local fractal dimension. First of all, convert the image in RGB color space into YCBCR. Secondly, original triangle prism model algorithm which calculates the square of block matrix is replaced with four pyramid model. This algorithm is easy to be realized, reduce the amount of calculation model, and speed up the image segmentation speed. Segmentation effect significantly is improved, and the simulation experimental results show that this algorithm can be very good for the segmentation image of the flame extracted from complex background. It has obvious, accurate and fast effect and provides a good foundation for the future image analysis and recognition processing.

Abstract: The distorted X-Ray Image Intensifier (XRII) image can introduce negative effect on following work for C-arm CT imaging system. In this paper, we propose an integrated approach based on least squares and Biharmonic spline interpolation to correct geometric distortions of XRII images. The method first uses morphology operation to extract the coordinate values of control points. Then the least square method fits the extracted coordinate values in every row and computes the more coordinate values by fixing the length in every row. Finally, The Biharmonic spline interpolation is used to interpolate the all coordinate values and correct the distortional XRII image. The experiment shows that the integrated method can effectively correct the distorted XRII image.

Abstract: This paper was proposed a new algorithm for Facial Expression Recognition (FER) which was based on fusion of gabor texture features and Centre Binary Pattern (CBP). Firstly, gabor texture feature were extracted from every expression image. Five scales and eight orientations of gabor wavelet filters were used to extract gabor texture features. Then the CBP features were extracted from gabor feature images and adaboost algorithm was used to select final features from CBP feature images. Finally, we obtain expression recognition results on the final expression features by Sparse Representation-based Classification (SRC) method. The experiment results on Japanese Female Facial Expression (JAFFE) database demonstrated that the new algorithm had a much higher recognition rate than the traditional algorithms.

Abstract: Analyses of GPS signals by wavelet algorithms and empirical mode decomposition (EMD) have demonstrated the strength of these techniques in discriminating signals from noise. However, the denoising precision seriously affects the final EMD error, especially for signals containing incremental developments in information. We present a new noise filter and trend extraction model based on the orthogonal wavelet transform and EMD. Simulated and real data are used to evaluate the proposed method. The results suggest that: 1) The orthogonal wavelet transform and EMD method can better mitigate the random errors hidden in periodic signals; 2) For signals with a linear trend, the orthogonal wavelet transform filtering method is superior to EMD. We suggest a method of trend extraction by EMD after noise filtering using the wavelet; 3) For signals with a nonlinear trend, theoretical analysis and simulation results show that the new noise filter and trend extraction model is superior to EMD and the simple combination of wavelets with EMD. The proposed approach not only extracts instantaneous features, but also reduces the number of decomposition layers of the signals and the cumulative errors in later decomposition. This method significantly improves the accuracy of the extracted deformation; 4) After mitigating the influence of multipath and other error effects with the new model, we attain millimeter accuracy for the vertical component position in GPS dynamic deformation.

Abstract: Proposes a new scheme for low quality fingerprint images which is used point oriental image and based on gray distributing rule of the pixels after investigating existing approaches to fingerprint segmentation. Experiment results indicated that this scheme performs better than traditional fingerprint image segmentation alogrithms. And it has higher performance in terms of efficiency and robustness.

Abstract: In the paper, degradation and restoration model is introduced. Image restoration method using inverse filtering and using wiener filtering are studied and implemented. A new method of image restoration is proposed by combining histogram equalization and median filtering. Comparing three methods by MATLAB simulation, the results show that the new method can effectively restore degradation image with comparatively high restoration efficiency.

Abstract: To solve the problem of translation compensation on ballistic target under complex scattering model in midcourse, a novel rule based on cycle subtracting of micro-Doppler curve is proposed. Firstly, the period is estimated by the period feature extraction algorithm which is based on the correlation of range profile envelopes. Then, extract the Doppler curve of the strongest scattering point through the Viterbi algorithm. Lastly, do the cycle subtracting and single cycle integral to the curve, the precious translation parameters can be got by the least squares estimating. This method is not only suitable for the rotating model, but also valuable for the precession model and the sliding model. The effectiveness of the proposed rule is verified by simulation results.

Abstract: Aiming at the illumination change and partial occlusion in the object tracking, an object tracking method based on illumination compensation was proposed. An illumination compensation method based on Retinex was applied to the sequence images, a structural appearance model and template matching were used to track the object. Dense sampling was used to obtain candidates, extended least median square was used to match templates, and a step by step template updating method is applied. The experimental results demonstrate the effect of the proposed method.

Abstract: This paper presents a technique for monocular Structure from Motion (SFM) that reconstructs 3D world shape. The technique proposed uses optical flow for 2D pixel pair matching and Angular Bundle Ajustment (ABA) for 3D structure refinement. The proposed strategy has two main advantages. Firstly, optical flow fields provide sufficient dense correspondence of image point pairs and secondly, ABA outperforms classic BA variants, especially for the points relatively far from camera. The reconstruction results obtained in realistic scenario demonstrate the effectiveness and accuracy of the proposed algorithm.