Chessboard size

[zhangyuran-gg]

Hello, can you share a picture of the chess board grid so that we can make a calibration board for testing. By the way, can the difference between the checkerboard and the back panel (the calibration board attached) be ignored (the calibration board is slightly larger than the checkerboard)?

[JianKangEgon]

which tool do you want to try? Lidar & camera calibration or just for camera intrinsic? It differs a little bit

[zhangyuran-gg]

您想尝试哪种工具？激光雷达和摄像头校准还是仅用于摄像头固有？它略有不同
Lidar & camera calibration. Thank you. By the way, what does the content in calibration_front.yaml file represent

[JianKangEgon]

Please kindly check the README to see the picture of the chessboard grid. The slight difference between the panel and board does not matter so much, cause there is the last step of visual validation to modify your result. As for the calibration_front file, it's just an intermediate product. You can check the tf_front.yaml for final info

[zhangyuran-gg]

Thank you for the restoration. By the way, after calculating the tf of lidar and lidar cameras, can we obtain the tf of camera_car？

[zhangyuran-gg]

请查看 README 以查看棋盘网格的图片。面板和板子之间的细微差异并不那么重要，因为还有最后一步的视觉验证来修改您的结果。至于 calibration_front 文件，它只是一个中间产品。您可以查看 tf_front.yaml 以获取最终信息

Hello, is the order of euler_angle in tf_front.yaml yaw, pitch, roll?

[JianKangEgon]

请查看 README 以查看棋盘网格的图片。面板和板子之间的细微差异并不那么重要，因为还有最后一步的视觉验证来修改您的结果。至于 calibration_front 文件，它只是一个中间产品。您可以查看 tf_front.yaml 以获取最终信息

Hello, is the order of euler_angle in tf_front.yaml yaw, pitch, roll?

yes, in the caculate_r_t.cpp file, line 326 ~ 327:

    Eigen::Quaterniond q_final(q[0], q[1], q[2], q[3]);
    Eigen::Vector3d euler_angle = q_final.toRotationMatrix().eulerAngles(2, 1, 0);

show that the final output follows the order yaw, pitch, roll

[JianKangEgon]

Thank you for the restoration. By the way, after calculating the tf of lidar and lidar cameras, can we obtain the tf of camera_car？

What do you mean the tf of camera_car? This relative position depends on where you mount the camera on the car.

[zhangyuran-gg]

请查看 README 以查看棋盘网格的图片。面板和板子之间的细微差异并不那么重要，因为还有最后一步的视觉验证来修改您的结果。至于 calibration_front 文件，它只是一个中间产品。您可以查看 tf_front.yaml 以获取最终信息

您好，tf_front.yaml yaw、pitch、roll 中的euler_angle顺序是吗？

是的，在 caculate_r_t.cpp 文件中，第 326 ~ 327 行：

    Eigen::Quaterniond q_final(q[0], q[1], q[2], q[3]);
    Eigen::Vector3d euler_angle = q_final.toRotationMatrix().eulerAngles(2, 1, 0);

显示最终输出遵循 yaw、pitch、roll 的顺序

Thank you for your reply. The program ran very well in the ros melodic version. I think there were some issues when running the program in the ros noetic and pcl-1.10 environment. The problem is that the viewer cannot display the point cloud. I hope to get your help, thank you very much

[JianKangEgon]

请查看 README 以查看棋盘网格的图片。面板和板子之间的细微差异并不那么重要，因为还有最后一步的视觉验证来修改您的结果。至于 calibration_front 文件，它只是一个中间产品。您可以查看 tf_front.yaml 以获取最终信息

您好，tf_front.yaml yaw、pitch、roll 中的euler_angle顺序是吗？

是的，在 caculate_r_t.cpp 文件中，第 326 ~ 327 行：

    Eigen::Quaterniond q_final(q[0], q[1], q[2], q[3]);
    Eigen::Vector3d euler_angle = q_final.toRotationMatrix().eulerAngles(2, 1, 0);

显示最终输出遵循 yaw、pitch、roll 的顺序

Thank you for your reply. The program ran very well in the ros melodic version. I think there were some issues when running the program in the ros noetic and pcl-1.10 environment. The problem is that the viewer cannot display the point cloud. I hope to get your help, thank you very much

Sorry, there is no plan to make it compatible in different environments. This kind of issue may mainly caused by version changes, you could focus on the difference between APIs to solve the problem.

[zhangyuran-gg]

请查看 README 以查看棋盘网格的图片。面板和板子之间的细微差异并不那么重要，因为还有最后一步的视觉验证来修改您的结果。至于 calibration_front 文件，它只是一个中间产品。您可以查看 tf_front.yaml 以获取最终信息

您好，tf_front.yaml yaw、pitch、roll 中的euler_angle顺序是吗？

是的，在 caculate_r_t.cpp 文件中，第 326 ~ 327 行：

    Eigen::Quaterniond q_final(q[0], q[1], q[2], q[3]);
    Eigen::Vector3d euler_angle = q_final.toRotationMatrix().eulerAngles(2, 1, 0);

显示最终输出遵循 yaw、pitch、roll 的顺序

Thank you for your reply. The program ran very well in the ros melodic version. I think there were some issues when running the program in the ros noetic and pcl-1.10 environment. The problem is that the viewer cannot display the point cloud. I hope to get your help, thank you very much

Sorry, there is no plan to make it compatible in different environments. This kind of issue may mainly caused by version changes, you could focus on the difference between APIs to solve the problem.

Using my three sets of data for calculation, the visualization result shows a small projection error, but the tf_front.yaml result is slightly different from the ideal value. Why is this? Perhaps I should test with more sets of data, tf results [-2.5, -0.165, -0.499, 3.11, -0.08, -1.11], ideal values [-1.82, 0.02, 0.05, 3.141, 0, -1.57]