ORB_SLAM2中标定数据的问题

Posted on In , Views: Disqus:

在使用ORB_SLAM2的过程中，我使用Kinect v2作为摄像机，而在使用之前需要对Kinect进行标定的工作。幸好iai_kinect2这个驱动提供了标定的工具1。按照说明操作，获得了标定的数据，如calib_color.yaml文件中包含了摄像机的内参和畸变等参数：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
%YAML:1.0
cameraMatrix: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 1.0550860028898474e+03, 0., 9.7022756868552835e+02, 0.,
       1.0557186689448556e+03, 5.2645231780561619e+02, 0., 0., 1. ]
distortionCoefficients: !!opencv-matrix
   rows: 1
   cols: 5
   dt: d
   data: [ 5.0049307122037007e-02, -5.9715363588982606e-02,
       -1.6247803478461531e-03, -1.3650166721283822e-03,
       1.2513177850839602e-02 ]
rotation: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 1., 0., 0., 0., 1., 0., 0., 0., 1. ]
projection: !!opencv-matrix
   rows: 4
   cols: 4
   dt: d
   data: [ 1.0550860028898474e+03, 0., 9.7022756868552835e+02, 0., 0.,
       1.0557186689448556e+03, 5.2645231780561619e+02, 0., 0., 0., 1.,
       0., 0., 0., 0., 1. ]

然后根据下式，提取上面的数据，填到ORB_SLAM2定义的校正参数yaml文件中2：

cameraMatrix:

\[\left[ {\begin{array}{*{20}{c}} {f_x}&0&{c_x}\\ 0&{f_y}&{c_y}\\ 0&0&1 \end{array}} \right]\]

distortionCoefficients:

\[\left[ {\begin{array}{*{20}{c}} {k_1}&{k_2}&{p_1}&{p_2}&{k_3} \end{array}} \right]\]

然后yaml文件内容如下：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
%YAML:1.0
    #--------------------------------------------------------------------------------------------
    # Camera Parameters. Adjust them!
    #--------------------------------------------------------------------------------------------

    # Camera calibration and distortion parameters (OpenCV) 
    Camera.fx: 1.0550860028898474e+03
    Camera.fy: 1.0557186689448556e+03
    Camera.cx: 9.7022756868552835e+02
    Camera.cy: 5.2645231780561619e+02

    Camera.k1: 5.0049307122037007e-02
    Camera.k2: -5.9715363588982606e-02
    Camera.p1: -1.6247803478461531e-03
    Camera.p2: -1.3650166721283822e-03
    Camera.k3: 1.2513177850839602e-02

    Camera.width: 960
    Camera.height: 540

    # Camera frames per second 
    Camera.fps: 30.0

    # Color order of the images (0: BGR, 1: RGB. It is ignored if images are grayscale)
    Camera.RGB: 1

    # IR projector baseline times fx (aprox.)
    Camera.bf: 40.0

    # Close/Far threshold. Baseline times.
    ThDepth: 50.0

    # Deptmap values factor 
    DepthMapFactor: 1000.0

    #--------------------------------------------------------------------------------------------
    # ORB Parameters
    #--------------------------------------------------------------------------------------------

    # ORB Extractor: Number of features per image
    ORBextractor.nFeatures: 1000

    # ORB Extractor: Scale factor between levels in the scale pyramid     
    ORBextractor.scaleFactor: 1.2

    # ORB Extractor: Number of levels in the scale pyramid    
    ORBextractor.nLevels: 8

    # ORB Extractor: Fast threshold
    # Image is divided in a grid. At each cell FAST are extracted imposing a minimum response.
    # Firstly we impose iniThFAST. If no corners are detected we impose a lower value minThFAST
    # You can lower these values if your images have low contrast            
    ORBextractor.iniThFAST: 20
    ORBextractor.minThFAST: 7

    #--------------------------------------------------------------------------------------------
    # Viewer Parameters
    #--------------------------------------------------------------------------------------------
    Viewer.KeyFrameSize: 0.05
    Viewer.KeyFrameLineWidth: 1
    Viewer.GraphLineWidth: 0.9
    Viewer.PointSize:2
    Viewer.CameraSize: 0.08
    Viewer.CameraLineWidth: 3
    Viewer.ViewpointX: 0
    Viewer.ViewpointY: -0.7
    Viewer.ViewpointZ: -1.8
    Viewer.ViewpointF: 500

在解决了双重标定的问题后3，我使用qhd质量(960x540)的图片跑ORB_SLAM2程序，发现无论是单目模式还是RGBD模式的结果都不堪理想。经过排查后，发现还是标定数据的问题。

在iai_kinect2的标定程序中，使用的FullHD(1920x1080)分辨率图片，所以得到的计算机内参数据是针对1920x1080这个分辨率的；而我在ORB_SLAM2中，使用的是QHD(960x540)分辨率的图片。为了使用标定数据与使用照片对应，需要对1920x1080下的标定数据作出处理，对内参数据根据分辨率按比例进行缩减4，在这里，需要对\(f_x\), \(f_y\), \(c_x\), \(c_y\)的值乘以一个0.5。

最终yaml文件内容如下：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
%YAML:1.0
    #--------------------------------------------------------------------------------------------
    # Camera Parameters. Adjust them!
    #--------------------------------------------------------------------------------------------

    # Camera calibration and distortion parameters (OpenCV) 
    Camera.fx: 527.54300144
    Camera.fy: 527.85933447
    Camera.cx: 485.11378434
    Camera.cy: 263.2261589

    Camera.k1: 5.0049307122037007e-02
    Camera.k2: -5.9715363588982606e-02
    Camera.p1: -1.6247803478461531e-03
    Camera.p2: -1.3650166721283822e-03
    Camera.k3: 1.2513177850839602e-02
    
    ...

重新运行ORB_SLAM2，问题解决。

  1. https://github.com/code-iai/iai_kinect2/tree/master/kinect2_calibration↩︎

  2. http://www.jianshu.com/p/c3e8c88edb64↩︎

  3. http://luohanjie.com/2017-03-30/dual-calibration-problem-in-orb-slam2-and-iai-kinect2.html↩︎

  4. http://dsp.stackexchange.com/questions/6055/how-does-resizing-an-image-affect-the-intrinsic-camera-matrix/6057#6057↩︎