Fangcheng Zhu
Fangcheng Zhu
> > Getting different results each time is normal, as any algorithm calibration comes with errors. If the differences in your results are not significant (such as a rotation of...
> 您好，麻烦您几个问题 1.Online Refinement阶段是继续做旋转激励（在原地像视频中一样旋转），还是做一些平移运动比较好（比如跑一个完整的回环指令） 2.外参数校准结果、特别是平移与设定的Online Refinement有很大关系，比如有时候我优化10s结果很好，而优化30s结果就很差，或者我有其他方法判断外参收敛到准确值吗 online_refinement阶段建议是继续沿三个轴画弧线，保证同时有旋转和平移激励（像视频中一样）。refine阶段实际上运行的是FAST-LIO2，其需要优化的状态量很多，所以不能保证外参标定精度有多高。我的经验和建议是多标定几次，计算几次initialization_result的均值作为最终的外参结果。
> what It seems strange. If a point cloud frame is splitted into 10 parts based on timestamps, each frame would only have a FoV of 36°, which could easily...
It seems that the device is car-like of which no sufficient excitation can be guaranteed. Current version does not support car-like device.
> rosbag This issue may be caused by an excessive number of points and the non-real-time lidar odometry. Due to the non-real-time case of LO, the IMU buffer stores a...
>  [  ](https://private-user-images.githubusercontent.com/38457117/299621509-527ed5f3-dc55-46a0-b966-38dcb354fc2a.gif?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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._Ix42yl91653bZMaeRgZTyXkL_pY1H1BaVnx_u-A8fE) [ ](https://private-user-images.githubusercontent.com/38457117/299621509-527ed5f3-dc55-46a0-b966-38dcb354fc2a.gif?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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._Ix42yl91653bZMaeRgZTyXkL_pY1H1BaVnx_u-A8fE) > > Hi! Thanks for great work!! > > I ran the code in my setup using Velodyne 32 channels and found that...
> Thank you for your kind reply. It seems that I should use the result of initialization. > > However, the recommended value for cut_frame_num = 3 ~ 4 is...
> @zfc-zfc 您好！感谢优秀的工作。 我按照您在 #19 中的提示， 测试激光里程计时。发现对于机械式雷达，FAST-LO表现的都很好，但针对于固态雷达如mid-360或avia，效果则变得很不稳定，经常出现抖动的现象。为了排除我自己数据集的影响，我也下载了您上传的mid360_100ms_delay.bag进行了测试。 > > 使用的参数如下。 > > ``` > common: > lid_topic: "/livox/lidar" > imu_topic: "/livox/imu/async" > > preprocess: > lidar_type: 1 # Livox series...
> 首先非常感谢大佬的工作。 在尝试进行HAP激光雷达和外置IMU的标定时，似乎平移上存在一些问题，我们标定的场景是室内场景。 这是我的角速度和线加速度结果的截图，麻烦您帮我看看存在什么问题，非常感谢！   看右下方这张图片，XY方向的加速度激励可能太小了，导致约束不足
> 那可以请问一下，为什么左下方图片中激光雷达里程计和IMU的X Y轴的加速度差别这么大吗？（可以明显看出一开始IMU的加速度的变化范围要比激光雷达里程计大） 加速度的变化范围不同是很正常的。刚体上任意两点AB的加速度与角速度关系是不同的，角速度满足 R\omega_A = \omega_B，所以AB两点的角速度norm相等；但加速度不是，AB两点的加速度关系可以参考论文公式(21)，由于有角速度、角加速度和平移外参的存在，会导致AB的加速度norm不同。