Font Size: a A A

The Research On Slope Deformation Of Terrestrial Three-dimensional Laser Scanning Point Cloud Data Based On Wavelet Neural Network

Posted on:2016-03-08Degree:MasterType:Thesis
Country:ChinaCandidate:L DanFull Text:PDF
GTID:2180330476951335Subject:Geodesy and Survey Engineering
Abstract/Summary:PDF Full Text Request
With the development of science and technology, traditional surveying and mapping technology cannot satisfy the demands of modern people to obtain the urgent requirement of the massive measured data. In terms of obtaining planar data, terrestrial three dimensional laser scanner has played a growing important role in collecting data.Because of its unchangeable status, extensive use demands the higher accuracy.because it’s hard to make a breakthrough on hardware, so we put the goal of improving precision of ground subsidence.Given that many scholars have used the time series and kalman filtering method to calculate the landslide area of subsidence, this paper uses the wavelet neural network analysis method accurately to calculate subsidence value of landslide area. The research content mainly includes the ground three-dimensional laser scanning data preprocessing part and the analysis of the wavelet neural network method for network training and ultimately obtained accurate settlement value of two parts.Terrestrial three-dimensional laser scanning points preprocessing contain joining together of point cloud, reduction of point cloud, resampling of point cloud and filtering of point cloud,which aim at improving the property of point cloud.Based on wavelet neural analysing the curved surface, and acquiring the deformation value.In the first place,trained surface with the point cloud data in the neural network for nonlinear countless approximation,Therefore an approximation to the point cloud data of implicit surface will come out. In the meanwhile, using BP neural network to surface reconstruction with point cloud data has been carried out. By comparing the two kinds of curved surface fitting precision of network model, wavelet neural network has advantages in the comparsion of BP neural network based on model fitting precision higher ground surface model. Finally,verifying the wavelet neural network is used to analyse the surface fitting the reliability of this method with the processor of GUI,its reliability were confirmed based on MATLAB platform,verifying the dependability and scientificity by way of using wavelet neural network to analyse the surface fitting.
Keywords/Search Tags:terrestrial three-dimensional laser scanning, point cloud data, wavelet neural network, surface reconstruction, slope monitoring
PDF Full Text Request
Related items