structure-from-motion revisited论文笔记

本文深入探讨了Structure-from-Motion(SfM)技术,包括对应搜索、增量式三维重建及其挑战。文章提出了增强的几何验证策略,改进的Next Best View Selection方法,更鲁棒的三角化技术和优化的Bundle Adjustment,显著提升了SfM的完整性、鲁棒性和准确性。通过对场景图的增强和多分辨率分析,确保了更高效的图片选择和三角测量,为大规模网络图片重建提供了高效解决方案。
摘要由CSDN通过智能技术生成

1. Introduction

提出一种新的SFM技术

SFM主要分为三类:增量式(most popular)\分层式\全局式;

但是目前在鲁棒性\精确度\完整性\扩展性方面还需要提高

2. Review of Structure-from-Motion

2.1 Correspondence Search---对应搜索

       (1)特征提取:SIFT

       (2)特征匹配:output:图片pair(a,b)、feature corrospondence(Mab = Fa × Fb)

       (3)几何验证:由于特征匹配都是基于特征点的apparence的相似度,无法保证某个特征点可以准确的map到same scene point。所以利用投影几何来估计一个transformation,来map图像对儿之间的特征点。

              · 具体方法如下:             

                     a.平面场景:单应矩阵(H)

                     b.非平面场景:利用对极几何---本质矩阵(F:未标定)、基本矩阵(E:标定)

              · 几何验证成功条件:如果transformation在图片对儿之间map的特征数量(内点,采用RANSAC方法)超过一定的阈值,就认为这两张图片通过了几何验证。

   · 几何验证的output:通过几何验证的图像pair、内点对儿(Mab = Fa × Fb)、【可选】几何关系Gab

     即生成scene graph,节点:图片;边:通过几何验证的图片对---verified pairs of images

tips:(单应矩阵、本质矩阵、基本矩阵讲解及约束关系推导:

              http://www.cnblogs.com/youzx/p/6385513.html#wiz_toc_1)

2.2 增量式三维重建

input:scene graph

output:pose(Np个,注册成功的图片数)、points(Nx个,构成点云的三维点数)、

       (1)初始化

              在scene graph中选取两个dense location(with many overlapping cameras)---照片匹配数量多

       (2)图片注册

              通过PnP可以将新图片注册到当前的模型中(利用新加入图片与已注册图片之间的匹配点对儿&已生成的三维点---生成2D-3D对应,然后进行三角测量估计出出新加入图片的pose)

              排除外点的影响:

                     a.标定的:用RANCAS & minimal pose slover

                     b.未标定的:minimal slovers 或者 sampling-based approaches

                     本文提出一种next best image selection方法准确估计pose、可靠的三角化

       (3)三角化

              新注册的图片需要满足的条件:1.看到已经生成的三维点; 2.通过与已注册的图片之间的新匹配生成新的三维点;

              而三角化的作用,1.通过2D-3D对应实现新图片的注册;2.计算得到新的三维点

              本文提出一种robust&高效的三角化方法<

  • 9
    点赞
  • 21
    收藏
    觉得还不错? 一键收藏
  • 1
    评论
Bayesian Color Constancy is a computational framework that attempts to solve the problem of color constancy in computer vision. Color constancy refers to the ability of the human visual system to perceive the colors of objects as relatively constant, despite changes in the spectral composition of the illuminating light. The Bayesian Color Constancy framework is based on the idea that the visual system should use prior knowledge about the statistical properties of the colors in the scene, as well as the statistical properties of the illuminant, to infer the true colors of the objects in the scene. This is done by formulating a probabilistic model of the relationship between the colors of the objects and the illuminant, and then using Bayesian inference to estimate the most likely illuminant given the observed colors. The paper "Bayesian Color Constancy Revisited" presents a review of the Bayesian Color Constancy framework, and proposes several improvements to the original model. One of the main contributions of the paper is the development of a new method for estimating the prior distribution of illuminants, based on a statistical analysis of the colors in a large set of natural images. The authors also propose several modifications to the likelihood function used in the model, in order to better capture the properties of real-world images. Overall, the paper provides a comprehensive overview of the Bayesian Color Constancy framework, and presents several important improvements and extensions to the original model. These developments have the potential to improve the accuracy and robustness of color constancy algorithms in computer vision applications.
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值