Combines the wavelet analysis and neural network, this paper will be processed the data and the traditional BP neural network and kalman filter are analyzed and compared. First of all to obtain data of dam deformation wavelet denoising, excluding the contaminated data, obtain the optimal data set. Threshold denoising is generally adopted. Then based on the BP neural network, wavelet analysis to improve the traditional neural network model. Improve the underlying layer upon layer number and the number of nodes. Combined with the optimized dam deformation data, using the improved network model, the results to the regression model, ordinary kalman filter, this paper compares and analyzes the prediction effect evaluation.Comparison result is more ideal, which indicates that the combination of wavelet neural network model for deformation data processing has a good precision.
The UAV system acts as one of the infrastructure of earth observation, with its mobility, high speed, flexibility, economy and other remarkable technical advantages, has been widely used in various fields of the national economic construction, such as agricultural monitoring, resource development, disaster emergency treatment. Taking an actual engineering as a case study in this paper, the method and the skill of making digital orthophoto map was stated by using the UASMaster, the professional UAV data processing software, based on the eBee unmanned aerial vehicle. Finally, the precision of the DOM was analyzed in detail through two methods, overlapping the DOM with the existing DLG of the region and contrasting the points of the existing DLG of 1:1000 scale with the corresponding checkpoints of the stereomodel.
Deformation monitoring data of high-speed railway bridges will inevitably be affected because of noise pollution, A deformation monitoring point of high-speed railway bridge was measurd by using sokkia SDL30 electronic level for a long time,which got a large number of deformation monitoring data. Based on the characteristics of the deformation monitoring data of high-speed railway bridge, which contain lots of noise. Based on the MATLAB software platform, 120 groups of deformation monitoring data were applied to analysis of wavelet denoising.sym6,db6 wavelet basis function were selected to analyze and remove the noise.The original signal was broken into three layers wavelet,which contain high frequency coefficients and low frequency coefficients.However, high frequency coefficient have plenty of noise.Adaptive method of soft and hard threshold were used to handle in the high frequency coefficient.Then,high frequency coefficient that was removed much of noise combined with low frequency coefficient to reconstitute and obtain reconstruction wavelet signal.Root Mean Square Error (RMSE) and Signal-To-Noise Ratio (SNR) were regarded as evaluation index of denoising,The smaller the root mean square error and the greater signal-to-noise ratio indicate that them have a good effect in denoising.
We can surely draw some conclusions in the experimental analysis:the db6 wavelet basis function has a good effect in wavelet denoising by using a adaptive soft threshold method,which root mean square error is minimum and signal-to-noise ratio is maximum.Moreover,the reconstructed image are more smooth than original signal denoising after wavelet denoising, which removed noise and useful signal are obtained in the original signal.Compared to the other three methods, this method has a good effect in denoising, which not only retain useful signal in the original signal, but aiso reach the goal of removing noise. So, it has a strong practical value in a actual deformation monitoring project.
With the advantage of high speed, big transport capacity, low energy consumption, good economic benefits and so on, high-speed railway is becoming more and more popular all over the world. It can reach 350 kilometers per hour, which requires high security performances. So research on the prediction of high-speed railway settlement that as one of the important factors affecting the safety of high-speed railway becomes particularly important. This paper takes advantage of genetic algorithms to seek all the data in order to calculate the best result and combines the advantage of strong learning ability and high accuracy of wavelet neural network, then build the model of genetic wavelet neural network for the prediction of high-speed railway settlement. By the experiment of back propagation neural network, wavelet neural network and genetic wavelet neural network, it shows that the absolute value of residual errors in the prediction of high-speed railway settlement based on genetic algorithm is the smallest, which proves that genetic wavelet neural network is better than the other two methods. The correlation coefficient of predicted and observed value is 99.9%. Furthermore, the maximum absolute value of residual error, minimum absolute value of residual error-mean value of relative error and value of root mean squared error(RMSE) that predicted by genetic wavelet neural network are all smaller than the other two methods’. The genetic wavelet neural network in the prediction of high-speed railway settlement is more stable in terms of stability and more accurate in the perspective of accuracy.
Using the eleven Radiosonde Stations’ data in southwestern of China from 2010 to 2013 to calculate the conversion coefficient K which is a reference value of Precipitable Water Vapor (PWV). Then build the EMARDSON model and the EMARDSON K model which introduced with elevation parameter and altitude. And to analysis the accuracy of the two models in the southwest China by radiosonde data in 2014. The results show: 1) The K value calculated by EMARDSON model has good adaptability in southwest region. 2) The method of spatial interpolation prediction by choosing 7 Radiosonde Stations’ K value uniformity is more adaptable than using 11 Radiosonde Stations’ K value to build basic model in the case of predicting 11 Radiosonde Stations’ K value, and it has a certain accuracy when predicting by using spatial interpolation in some areas where lacking data. 3) The accuracy by using the A-EMARDSON model to predict K value was improved obviously. at the same time, when predicting K value by the method of spatial interpolation, both the precision of inner and the precision of outer are better than EMARDSON model. So it can be concluded that the altitude factor is an important factor to influence the K value prediction.