支持向量学习机在点云去噪中的应用  

Application of Support Vector Machine in Point Clouds Denoising

作　　者：张琴[1] 蔡勇[2] 

机构地区：[1]西南科技大学计算机科学与技术学院,四川绵阳621000 [2]西南科技大学制造科学与工程学院,四川绵阳621000

出　　处：《计算机技术与发展》2011年第6期85-88,94,共5页Computer Technology and Development

基　　金：国家自然科学基金项目(10576027)

摘　　要：在解决非线性、高维模式识别以及小样本等问题中,支持向量机表现出许多独有的优势。提出将支持向量机学习分类方法应用于点云去噪中,能够稳定地进行机器学习,训练得到判别模型,快速、准确地识别出噪声点与非噪声点。通过对小样本数据的统计学习,能够推广到大规模数据中去进行结果的预测估计。用SVM对点云数据样本进行学习训练、测试,识别分类,从而达到去噪光顺的目的。实验表明,此方法在有效去除噪声的同时能较完整地保留点云数据信息。SVM represented many unique advantages in many applications,such as solving the problem of nonlinear,high dimension pattern recognition and small sample problem.A method was promoted in this paper.It was removing point clouds using SVM sorting technique.It could be generalized to test and estimate the big scale data by training the small sample.SVM classification method was used to train,test,classify the point clouds data sample,so that it could achieve to the goal of denoising.The experiment showed that this method could remove the noise effectively while preserving the point clouds information relative completely.

关 键 词：点云 支持向量机 去噪 分类 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]