|国家预印本平台
首页|基于多尺度分析的截面线特征点自动提取及B样条曲线拟合

基于多尺度分析的截面线特征点自动提取及B样条曲线拟合

B-spline Curve Fitting Based on Multi-scale Coexist Automatic Recognition Feature Point

中文摘要英文摘要

针对多尺度特征并存截面数据的曲线拟合,提出一种尺度分析与去除细节小波重建相结合实现截面线去噪和特征点自动识别方法,并在此基础上实现截面线的拟合。首先根据截面线局部形状尺度因子,将截面线分成平滑段和波动段,针对不同尺度的特征段在相应的尺度下分别对各段用去除细节小波重建去除噪声,然后对去噪后的局部特征段根据截面线曲率、曲率线找出截面线上的显著点和拐点,构造插值于这些特征点的B样条曲线,就可以实现对原曲线的拟合。实验结果表明,所构造的曲线既能剔除噪声,又能更好的保留原截面线的特征。

ccording to a multi-scale coexist section data of curve fitting, put forward a kind of multiscale analysis and remove details combining wavelet reconstruction realize section line denoising and feature point to be automatic identification method, and on this basis to achieve cross-section line fitting. First, according to local shape scale factor section line, divide the cross section into smooth section and wave section. Characteristics at different scales in the corresponding section of the scale were used to remove the details section of the wavelet reconstruction to remove noise, and then to go after the noise the local characteristics of section according to section line curvature, find out the curvature line online significant point and cross section inflection point, the tectonic interpolation in these feature point B spline curve, can be achieved on the original curve fitting. The experimental results show that the constructed curve can eliminate noise, and better able to preserve the original section of line features.

李奇敏、尹小奎

计算技术、计算机技术工程设计、工程测绘数学

多尺度分析小波分析特征点曲线拟合

Multi-scale analysisWavelet analysisFeature pointsCurve fitting

李奇敏,尹小奎.基于多尺度分析的截面线特征点自动提取及B样条曲线拟合[EB/OL].(2012-06-21)[2025-08-18].http://www.paper.edu.cn/releasepaper/content/201206-303.点此复制

评论