資源簡介
SIFT特征是基于物體上的一些局部外觀的興趣點而與影像的大小和旋轉無關。對于光線、噪聲、微視角改變的容忍度也相當高。基于這些特性,它們是高度顯著而且相對容易擷取,在母數龐大的特征數據庫中,很容易辨識物體而且鮮有誤認。使用SIFT特征描述對于部分物體遮蔽的偵測率也相當高,甚至只需要3個以上的SIFT物體特征就足以計算出位置與方位。在現今的電腦硬件速度下和小型的特征數據庫條件下,辨識速度可接近即時運算。SIFT特征的信息量大,適合在海量數據庫中快速準確匹配。

代碼片段和文件信息
%?im?=?appendimages(image1?image2)
%
%?Return?a?new?image?that?appends?the?two?images?side-by-side.
function?im?=?appendimages(image1?image2)
%?Select?the?image?with?the?fewest?rows?and?fill?in?enough?empty?rows
%???to?make?it?the?same?height?as?the?other?image.
rows1?=?size(image11);
rows2?=?size(image21);
if?(rows1??????image1(rows21)?=?0;
else
?????image2(rows11)?=?0;
end
%?Now?append?both?images?side-by-side.
im?=?[image1?image2];???
?屬性????????????大小?????日期????時間???名稱
-----------?---------??----------?-----??----
?????目錄???????????0??2005-07-08?04:54??siftDemoV4\
?????文件????????1920??2005-07-08?04:54??siftDemoV4\defs.h
?????文件????????4876??2005-07-08?04:54??siftDemoV4\match.c
?????文件?????????631??2005-07-08?04:54??siftDemoV4\Makefile
?????文件????????7952??2005-07-08?04:54??siftDemoV4\README
?????文件???????97697??2005-07-08?04:54??siftDemoV4\book.pgm
?????文件??????196623??2005-07-08?04:54??siftDemoV4\scene.pgm
?????文件????????1940??2005-07-08?04:54??siftDemoV4\match.m
?????文件?????????461??2005-07-08?04:54??siftDemoV4\appendimages.m
?????文件????????1699??2005-07-08?04:54??siftDemoV4\showkeys.m
?????文件????????2494??2005-07-08?04:54??siftDemoV4\sift.m
?????文件????????7493??2005-07-08?04:54??siftDemoV4\util.c
?????文件???????45070??2005-07-08?04:54??siftDemoV4\sift
?????文件????????1055??2005-07-08?04:54??siftDemoV4\LICENSE
?????文件???????46433??2005-07-08?04:54??siftDemoV4\basmati.pgm
?????文件???????72310??2005-07-08?04:54??siftDemoV4\box.pgm
?????文件???????94208??2005-07-08?04:54??siftDemoV4\siftWin32.exe
- 上一篇:海思rtp封包發送
- 下一篇:中國聯通光纖分布系統網管技術規范V2.0
評論
共有 條評論