opencv 2.4 image stitching

 

首先介紹環境架設，可以參考http://www.cnblogs.com/freedomshe/archive/2012/04/25/2470540.html，

需要再新增的路徑與檔案如下：

STEP5：

Include目錄：

• C:\OpenCV2.4beta\modules\stitching\include\opencv2
• C:\OpenCV2.4beta\modules\stitching\include

STEP6：

其他相依性：

• opencv_stitching240d.lib

範例程式來源：C:\OpenCV2.4beta\samples\cpp\stitching.cpp

接著我都會Win + R開啟命令提示字元，然後到專案執行檔的資料夾路徑，再把要處理的圖都進去，接著以指令如：

• StitchTest.exe a.jpg b.jpg c.jpg

a.jpg b.jpg c.jpg三張圖就會兩兩特徵比對，也就是說順序無關，若有更多影像要拼接也可繼續接在其後。

需等待一段時間才有結果，過程陸續會有訊息出現：

Finding features…
Features in image #1: 1250
Features in image #2: 1744
Features in image #3: 1441
Finding features, time: 11.8387 sec
Pairwise matching
1->2 matches: 9

1->2 matches: 80

1->2 matches: 118
1->2 & 2->1 matches: 162
.1->2 & 2->1 matches: 51
1->2 & 2->1 matches: 136
..Pairwise matching, time: 145.144 sec
Estimating rotations…
Estimating rotations, time: 0.0145659 sec
Initial intrinsic parameters #1:
[466.2391772970087, 0, 400;
0, 466.2391772970087, 267;
0, 0, 1]
Initial intrinsic parameters #2:
[466.2391772970087, 0, 400;
0, 466.2391772970087, 267;
0, 0, 1]
Initial intrinsic parameters #3:
[466.2391772970087, 0, 400;
0, 466.2391772970087, 267;
0, 0, 1]
Camera #1:
[737.0802894092269, 0, 400;
0, 737.0802894092269, 267;
0, 0, 1]
Camera #2:
[735.834675960764, 0, 400;
0, 735.834675960764, 267;
0, 0, 1]
Camera #3:
[742.896895252395, 0, 400;
0, 742.896895252395, 267;
0, 0, 1]
Wave correcting…
Wave correcting, time: 0.0184775 sec
Warping images (auxiliary)…
Warping images, time: 0.259227 sec
Exposure compensation…
Exposure compensation, time: 0.317293 sec
Finding seams…
Finding seams, time: 0.561857 sec
Compositing…
Compositing image #1
Compositing image #2
Compositing image #3
Compositing, time: 1.65429 sec

可見拼接的步驟相當多，而且所花的時間總和也有點久，這時候我們會好奇，OpenCV的Stitching有哪些步驟，這可以看官方網站的Stitching Pipeline：（可點圖放大）

以上步驟可視需求調整，來加快拼接速度。pairwise matching花比較多的時間，因為兩兩比對特徵：

接下來看實驗結果如何: