目录

1 安装环境

1.1 启动相机和标定

2 准备棋盘格

2.1 打印一张棋盘格图片

2.2 具体标定过程

2.3 所有结果与出厂参数对比汇总

参考

---

1 安装环境

sudo apt install ros-humble-camera-calibration

1.1 启动相机和标定

ros2 launch realsense2_camera rs_launch.py

（1）--size 9x6： 为当前标定板的大小（棋盘格是内部小正方形角点的交点！）

（2）--square 0.026：为每个小棋盘格的边长，单位是米

（3）image:=/camera/image_raw：标定当前订阅图像来源自名为/camera/image_raw的topic （4）camera:=/camera：为摄像机名

ros2 run camera_calibration cameracalibrator --size 11x8 --square 0.02 --ros-args --remap image:=/camera/color/image_raw --remap camera:=/camera

9*6内角点棋盘格，格子大小为26mm:

ros2 run camera_calibration cameracalibrator --size 9x6 --square 0.026 --ros-args --remap image:=/camera/color/image_raw --remap camera:=/camera

2 准备棋盘格

2.1 打印一张棋盘格图片

可以从这里获取 ，下面给出opencv官网给出的一张9*6内角点棋盘格，格子大小为26mm：

opencv ——pattern.png

2.2 具体标定过程

8x6和0.108是官方给出的标定板的尺寸，请根据自己的修改比如我的是8x6和0.0216（参考）

为了得到一个好的标定结果，应该使得标定板尽量出现在摄像头视野的各个位置里：

如标定板出现在视野中的左边，右边，上边和下边，标定板既有倾斜的，也有水平的.所以多动一动～

界面中的

x：表示标定板在视野中的左右位置。

y：表示标定板在视野中的上下位置。

size：标定板在占视野的尺寸大小，也可以理解为标定板离摄像头的远近。

skew：标定板在视野中的倾斜位置。

因此，需要移动标定板：x代表左右移动，y代表上下移动，size代表远近移动，skew代表倾斜侧角，可以上下倾，也可以左右倾。

1、通过不停的移动标定板：直到X、Y、Size、Skew四个都变成绿色

2、当四个都变绿色时，下面的三个按钮也会变成青色，此时点击CALIBRATE开始标定，过程大约1,2分钟

当开始标定后，可以在终端中看到看到标定的信息：

*** Added sample 1, p_x = 0.533, p_y = 0.686, p_size = 0.443, skew = 0.007
*** Added sample 2, p_x = 0.432, p_y = 0.588, p_size = 0.442, skew = 0.022
*** Added sample 3, p_x = 0.411, p_y = 0.771, p_size = 0.446, skew = 0.000
*** Added sample 4, p_x = 0.417, p_y = 0.979, p_size = 0.446, skew = 0.004
*** Added sample 5, p_x = 0.446, p_y = 0.362, p_size = 0.437, skew = 0.020
*** Added sample 6, p_x = 0.451, p_y = 0.162, p_size = 0.438, skew = 0.013
*** Added sample 7, p_x = 0.472, p_y = 0.530, p_size = 0.340, skew = 0.033
*** Added sample 8, p_x = 0.538, p_y = 0.455, p_size = 0.306, skew = 0.009
*** Added sample 9, p_x = 0.475, p_y = 0.836, p_size = 0.496, skew = 0.095
*** Added sample 10, p_x = 0.556, p_y = 0.588, p_size = 0.485, skew = 0.064
*** Added sample 11, p_x = 0.569, p_y = 0.419, p_size = 0.473, skew = 0.083
*** Added sample 12, p_x = 0.596, p_y = 0.525, p_size = 0.417, skew = 0.107
*** Added sample 13, p_x = 0.546, p_y = 0.262, p_size = 0.417, skew = 0.038
*** Added sample 14, p_x = 0.543, p_y = 0.029, p_size = 0.423, skew = 0.069
*** Added sample 15, p_x = 0.371, p_y = 0.075, p_size = 0.369, skew = 0.071
*** Added sample 16, p_x = 0.375, p_y = 0.181, p_size = 0.335, skew = 0.010
*** Added sample 17, p_x = 0.379, p_y = 0.385, p_size = 0.334, skew = 0.008
*** Added sample 18, p_x = 0.358, p_y = 0.635, p_size = 0.364, skew = 0.001
*** Added sample 19, p_x = 0.382, p_y = 0.872, p_size = 0.374, skew = 0.001
*** Added sample 20, p_x = 0.568, p_y = 0.729, p_size = 0.486, skew = 0.137
*** Added sample 21, p_x = 0.603, p_y = 0.558, p_size = 0.488, skew = 0.230
*** Added sample 22, p_x = 0.638, p_y = 0.430, p_size = 0.499, skew = 0.279
*** Added sample 23, p_x = 0.658, p_y = 0.702, p_size = 0.435, skew = 0.060
*** Added sample 24, p_x = 0.665, p_y = 0.623, p_size = 0.407, skew = 0.150
*** Added sample 25, p_x = 0.690, p_y = 0.577, p_size = 0.424, skew = 0.025
*** Added sample 26, p_x = 0.484, p_y = 0.469, p_size = 0.434, skew = 0.143
*** Added sample 27, p_x = 0.454, p_y = 0.540, p_size = 0.399, skew = 0.210
*** Added sample 28, p_x = 0.451, p_y = 0.703, p_size = 0.402, skew = 0.159
*** Added sample 29, p_x = 0.375, p_y = 0.474, p_size = 0.411, skew = 0.047
*** Added sample 30, p_x = 0.374, p_y = 0.679, p_size = 0.443, skew = 0.073
*** Added sample 31, p_x = 0.667, p_y = 0.513, p_size = 0.308, skew = 0.022
*** Added sample 32, p_x = 0.782, p_y = 0.424, p_size = 0.308, skew = 0.028
*** Added sample 33, p_x = 0.695, p_y = 0.318, p_size = 0.308, skew = 0.045
*** Added sample 34, p_x = 0.390, p_y = 0.655, p_size = 0.327, skew = 0.203
*** Added sample 35, p_x = 0.399, p_y = 0.707, p_size = 0.331, skew = 0.342
*** Added sample 36, p_x = 0.497, p_y = 0.605, p_size = 0.336, skew = 0.128
*** Added sample 37, p_x = 0.510, p_y = 0.707, p_size = 0.336, skew = 0.244
*** Added sample 38, p_x = 0.497, p_y = 0.546, p_size = 0.339, skew = 0.306
*** Added sample 39, p_x = 0.635, p_y = 0.678, p_size = 0.310, skew = 0.006
*** Added sample 40, p_x = 0.611, p_y = 0.846, p_size = 0.308, skew = 0.012
*** Added sample 41, p_x = 0.577, p_y = 0.204, p_size = 0.307, skew = 0.030
**** Calibrating ****
mono pinhole calibration...
D = [0.13708451678437922, -0.25401772231120406, 0.009488680435853903, 0.0024256349577669247, 0.0]
K = [1342.207453796947, 0.0, 965.9585846877721, 0.0, 1341.2992664435826, 587.960370520978, 0.0, 0.0, 1.0]
R = [1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0]
P = [1347.8486328125, 0.0, 971.9469484728761, 0.0, 0.0, 1364.8638916015625, 594.9367710060869, 0.0, 0.0, 0.0, 1.0, 0.0]
None
# oST version 5.0 parameters


[image]

width
1920

height
1080

[narrow_stereo]

camera matrix
1342.207454 0.000000 965.958585
0.000000 1341.299266 587.960371
0.000000 0.000000 1.000000

distortion
0.137085 -0.254018 0.009489 0.002426 0.000000

rectification
1.000000 0.000000 0.000000
0.000000 1.000000 0.000000
0.000000 0.000000 1.000000

projection
1347.848633 0.000000 971.946948 0.000000
0.000000 1364.863892 594.936771 0.000000
0.000000 0.000000 1.000000 0.000000

3、然后点击Save按钮，会把标定的文件信息保存到：/tmp/calibrationdata.tar.gz路径下

calibrationdata.tar.gz压缩文件中，会存储标定过程中的图片，还有就是标定的参数相机参数文件，如下：

ost.txt文件内容：

# oST version 5.0 parameters


[image]

width
1920

height
1080

[narrow_stereo]

camera matrix
1346.337242 0.000000 973.784611
0.000000 1345.832807 554.702266
0.000000 0.000000 1.000000

distortion
0.145266 -0.265614 -0.000379 0.002002 0.000000

rectification
1.000000 0.000000 0.000000
0.000000 1.000000 0.000000
0.000000 0.000000 1.000000

projection
1353.022461 0.000000 979.819928 0.000000
0.000000 1373.117188 553.884378 0.000000
0.000000 0.000000 1.000000 0.000000

ost.yaml文件内容：

2.3 所有结果与出厂参数对比汇总

来源	fx	fy	cx	cy	k₁	k₂
出厂	1354.44	1352.06	970.9	559.16	0	0
第一次	1311.47	1310.83	954.27	553.21	0.103	−0.211
第二次	1349.42	1349.06	977.54	556.16	0.121	−0.227
第三次	1303.3	1299.84	981.83	585.41	0.128	−0.217
第四次	1347.47	1345.31	964.89	565.73	0.125	−0.243
第五次	1346.34	1345.83	973.78	554.7	0.145	−0.266

也可以用该方法标定红外深度摄像头

https://blog.csdn.net/hai411741962/article/details/133809244

参考

https://blog.csdn.net/xiaojiu1997/article/details/130607885

https://developer.aliyun.com/article/1599991

https://developer.aliyun.com/article/1592466