基于机器学习的主动式匹配算法
首发时间:2018-11-05
摘要:本文提出一种基于机器学习的主动式双目匹配算法,DOE(Diffraction Optical Element)装置作为主动光源,利用机器学习算法提取斑点图中的二值化特征向量,再利用特征向量进行汉明距离转换得到初始代价,最后对初始代价进行代价聚合和后处理,得到最终的视差值。该方法适用于一般场景,无需针对不同场景重复学习,算法的运行时间较传统的匹配算法不会有明显的增加。大量的实验结果表明,本文算法匹配整体效果好,精度也高于传统的匹配算法。
For information in English, please click here
Machine learning based active stereo match algorithm
Abstract:A active binocular stereo matching based on machine learning was proposed,DOE (Diffraction Optical Element) device as the active light source,using machine learning algorithms to extract the binary characteristic vector of the image,and then calculate the initial cost by converting the binary characteristic vector to the hamming distance,get the final disparity value.This method is applicable to general scenarios without repeated learning for different scenarios,and the running time of the algorithm is not significantly increased compared with the traditional matching algorithm.A large number of experimental results show that the algorithm in this paper has good overall matching effect and higher accuracy than traditional matching calculation.Extexnsive experimental results show that the algorithm in this paper has good overall matching effect and higher accuracy than traditional matching calculation.
Keywords: Machine Learning High Precision Stereo Match
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于机器学习的主动式匹配算法
评论
全部评论0/1000