Calibration process
Import the calibration data into the matlab lidar camara calibrator plug-in, click Import in the illustration and select Import Data as shown in the figure:
select the imported image and point cloud data in turn as follows and click " OK ":
Matlab will automatically import the data and calculate the internal parameters of the camera, and then Start to process the camera and point cloud data, and perform automatic calibration, but the automatic calibration results are generally poor, and it will prompt that no target is detected, which is a normal phenomenon.
Can be accurately calibrated after manual adjustment is required. After clicking " OK ", the adjustment process is as follows. First select Edit ROI, that is, to delineate the range of a checkerboard (calibration board). It does not need to be too large, as long as it can include all the sampled calibration boards, as shown in the figure.
Then adjust the area as shown below, and then click " Apply ": (It is still difficult to adjust the rotation and selection box, just try more)
Then adjust the Dimension Tolerance first , adjust it appropriately, and then select Select Checkerboard to select the point cloud of the calibration board. It is necessary to select the point cloud of each set of data .
Adjust a better posture, use the " small brush " in the picture to select the point cloud, and the frame can be selected. After the selection, the selected point cloud will turn red. Try to select only the point cloud of the calibration board , so that the calibration result is more accurate, as shown in the figure below.
Click " Apply " in the upper left corner and return to the plug-in home page, then click " Detect " and wait a while.
Then click " Calibrate "
The calibration results are as follows: It can be seen that the calibration board is covered with a blue point cloud, and the point cloud data is also assigned the color in the picture, including the color of the floor tiles. The lower left corner is the calibration error of each picture, and the middle is Pixel error, reconstruction error on the right. Then click " Export " to export the calibration parameters.
Then click OK .
Then you can open the result in the workspace of matlab. I don’t know what is wrong here. There will be a lot of calibration data before, including the internal parameters of the calibrated camera. . The variable T here is the extrinsic parameter matrix.
You can import the matrix on this website, and then you can see information such as the Euler angle between the camera and the radar. https://www.andre-gaschler.com/rotationconverter/ is as follows:
Camera internal reference calibration is similar to this one. The plug-in used is Camara Calitrator . After the obtained results are exported, the following IntrinsicMatrix is the internal reference matrix.
