摘要
针对智能驾驶中图像拼接存在的匹配点冗余和拼接后图像局部扭曲的问题,提出了一种基于匹配点和薄板样条函数模型优化的视差图像拼接方法.首先,根据图像匹配点的分布位置构建稀疏矩阵;其次,通过网格约束匹配点数量消除冗余匹配点,缩短计算薄板样条函数模型的时间;最后,采用改进的薄板样条函数模型进行图像配准.实验结果表明,所提方法消除了匹配点冗余并改善了图像扭曲问题,具有一定的优越性和有效性.
To solve the redundancy of matching points and local distortion of image after mosaic in intelligent driving,a parallax image mosaic algorithm based on optimization of matching points and improvement of thin plate spline function model is proposed.First,a sparse matrix is constructed according to the distribution positions of image matching points.Second,the number of mesh constrained matching points is used to eliminate redundant matching points,which reduces the time of calculating thin plate spline function model.Finally,an improved thin plate spline function model is used for image registration.The experimental results indicate that the proposed algorithm eliminates the redundancy of matching points and improves the image distortion problem,which has certain superiority and effectiveness.
作者
陈洁芳
黄昶
CHEN Jiefang;HUANG Chang(School of Communication and Electronic Engineering,East China Normal University,Shanghai 200241,China)
出处
《华东师范大学学报(自然科学版)》
北大核心
2025年第4期15-27,共13页
Journal of East China Normal University(Natural Science)
关键词
智能驾驶
视差图像拼接
改进薄板样条函数模型
稀疏矩阵
匹配点
intelligent driving
parallax image stitching
improved thin plate splines function model
sparse matrix
matching points
