Search

In order to improve the accuracy of correlated noise (CN) model for distributed video coding (DVC), this paper proposes a novel distribution parameter fitting algorithm based on the minimum Euclidean distance. The presented method can obtain the final fitted distribution parameter by using the minimum Euclidean distance to compare the Laplace probability density function (PDF) with the PDF computed utilizing the actual residual frame data. Experiment results show that the proposed distribution parameter fitting algorithm can improve the rate-distortion (R-D) performance of DVC significantly.

Detecting the location offset of laser spot center method to monitor roadbed settlement is a kind of portable, easy to layout method. Detection laser spot center offset, we can use gravity, circle fitting, ellipse fitting and Gaussian curve fitting, etc method. Gravity algorithm is only involves basic arithmetic and low precise, and ellipse fitting is higher precise. Because of the angle between the optical axis and the target, the laser spot is more approximate as a ellipse. Using the ellipse fitting algorithm, Sub-pixel precise threshold segmentation and Canny edge extraction we obtained an ideal monitoring precise of the roadbed settlement.

To overcome disadvantages of traditional detection methods of wood flour mesh number, a mesh number recognition algorithm based on external rectangle fitting and morphological characteristics has been studied. It makes use of minimum external rectangle with the boundary points obtained by the preprocessing of microscopic images. The external rectangles length is calculated when the area is the smallest. The experimental results demonstrate that the proposed algorithm has a good fitting accuracy and meets producing demands.

In order to find a more accurate center wavelength detection fitting algorithm for fiber Bragg grating (FBG) reflection spectrum, Based on the theory of neural network, we designed different neural network to study and fit the spectral data which has been acquired from FBG.In this paper,we have studied BP,RBFand wavelet neural network structure,comparing the network training time and the result of fitting curves ,wavelet neural network is proved to be the best fitting algorithm.

The paper elaborates the principle of the GPS leveling and conventional secondary surface fitting method. It introduces dynamic weighted fitting method and MATLAB V4 interpolation fitting method. Finally, using an example, the paper makes the comparison and gets some practical significance conclusions: conventional secondary surface fitting method can be used if the area is small and its terrain is simple; the area's terrain is flat, it can use V4 interpolation and When the area's terrain is fluctuated, it can use dynamic weighted fitting method.

A mutli-model modeling method based on local model is given. The modeling idea is firstly to find some data matching with the current working point from vast historical system input-output datasets, and in this paper, we give a new method of choose data information based on similarity of vector which improve the accuracy of data greatly. Secondly to choose the weight and optimum bandwidth then develop a local model using local polynomial fitting algorithm. With the change of working points, multiple local models are built. The effectiveness of the proposed method is demonstrated by simulation results.

A new local fitting algorithm of implicit quadric surface is presented for space scattered point cloud. For the fitting objective function, the algorithm obtains ten groups coefficients of implicit quadric surface with ten different constrains and gets the final coefficients by taking arithmetic average of ten groups coefficients. Based on the curvature formula of implicit quadric surface, the point cloud data curvature can be calculated. The numerical example demonstrated the efficiency of curvature estimation algorithm.