Skip to content

Commit

Permalink
going to 7.5
Browse files Browse the repository at this point in the history
  • Loading branch information
gaoxiang12 committed Feb 27, 2021
1 parent 9fe183c commit 8196cf4
Show file tree
Hide file tree
Showing 7 changed files with 255 additions and 256 deletions.
2 changes: 1 addition & 1 deletion chapters-cn/vo1.tex
Original file line number Diff line number Diff line change
Expand Up @@ -120,7 +120,7 @@ \subsubsection{BRIEF描述子}

BRIEF是一种\textbf{二进制}描述子,其描述向量由许多个0和1组成,这里的0和1编码了关键点附近两个随机像素(比如$p$$q$)的大小关系:如果$p$$q$大,则取1,反之就取0。如果我们取了128个这样的$p,q$,最后就得到128维由0、1组成的向量\textsuperscript{\cite{calonder2010brief}}。BRIEF使用了随机选点的比较,速度非常快,而且由于使用了二进制表达,存储起来也十分方便,适用于实时的图像匹配。原始的BRIEF描述子不具有旋转不变性,因此在图像发生旋转时容易丢失。而ORB在FAST特征点提取阶段计算了关键点的方向,所以可以利用方向信息,计算了旋转之后的“Steer BRIEF”特征使ORB的描述子具有较好的旋转不变性。

由于考虑到了旋转和缩放,使得ORB在平移、旋转和缩放的变换下仍有良好的表现。同时,FAST和BREIF的组合也非常高效,使得ORB特征在实时SLAM中非常受欢迎。我们在\autoref{fig:ORB}~中展示了一张图像提取ORB之后的结果,下面来介绍如何在不同的图像之间进行特征匹配。
由于考虑到了旋转和缩放,使得ORB在平移、旋转和缩放的变换下仍有良好的表现。同时,FAST和BRIEF的组合也非常高效,使得ORB特征在实时SLAM中非常受欢迎。我们在\autoref{fig:ORB}~中展示了一张图像提取ORB之后的结果,下面来介绍如何在不同的图像之间进行特征匹配。

\begin{figure}[!htp]
\centering
Expand Down
507 changes: 253 additions & 254 deletions chapters/vo1.tex

Large diffs are not rendered by default.

2 changes: 1 addition & 1 deletion chapters/vo2.tex
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
% !Mode:: "TeX:UTF-8"
\chapter{Visual Odometry: Part 2}
\chapter{Visual Odometry: Part II}
\label{cpt:vo2}
\label{cpt:8}
\begin{mdframed}
Expand Down
Binary file modified resources/vo1/corner-flat-line.pdf
Binary file not shown.
Binary file modified resources/vo1/exp1-result.pdf
Binary file not shown.
Binary file modified resources/vo1/pyramid.pdf
Binary file not shown.
Binary file modified slambook-en.pdf
Binary file not shown.

0 comments on commit 8196cf4

Please sign in to comment.