FAST-LIVO2 icon indicating copy to clipboard operation
FAST-LIVO2 copied to clipboard
verified publisher icon hku-mars

速腾16线雷达+d435相机可以复现嘛？

Open c51-source opened this issue 4 months ago • 5 comments

我利用速腾16线雷达+d435相机去复现过程中轨迹是正确的，但是点云是白色的，偶尔会有一点点错乱的颜色是什么原因，利用message_filters做的软同步，标定方法是利用cam_lidar_calibration，得到rpyxyz在换算成外参的旋转矩阵，下面是我的标定结果以及运行算法的截图 process[visualise_results-1]: started with pid [15846] process[assess-2]: started with pid [15847] [INFO] [1753099156.564098074]: Importing samples from: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/poses.csv [INFO] [1753099156.564558913]: 10 samples imported [INFO] [1753099157.032299]: Starting visualise_results [INFO] [1753099157.035458]: Opening file at: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/calibration_2025-07-21_19-57-02.csv [INFO] [1753099157.036326]: Using degrees for rotation

Total number of samples: 44 roll | Num samples after filtering: 43 pitch | Num samples after filtering: 42 yaw | Num samples after filtering: 44 x | Num samples after filtering: 41 y | Num samples after filtering: 44 z | Num samples after filtering: 44

---- Calculating average reprojection error on 10 samples ---- 1/ 10 | dist= 0.699m, dimerr= 62.015mm | error: 14.597pix --> 5.516mm 2/ 10 | dist= 0.745m, dimerr= 64.334mm | error: 25.940pix --> 10.351mm 3/ 10 | dist= 0.740m, dimerr= 59.449mm | error: 34.377pix --> 14.200mm 4/ 10 | dist= 0.591m, dimerr= 41.913mm | error: 17.080pix --> 6.921mm 5/ 10 | dist= 1.111m, dimerr= 49.753mm | error: 50.072pix --> 31.171mm 6/ 10 | dist= 0.710m, dimerr= 37.961mm | error: 11.019pix --> 4.552mm 7/ 10 | dist= 0.687m, dimerr= 53.206mm | error: 12.545pix --> 4.547mm 8/ 10 | dist= 0.810m, dimerr= 32.256mm | error: 45.064pix --> 20.254mm 9/ 10 | dist= 0.663m, dimerr= 32.691mm | error: 31.299pix --> 10.704mm 10/ 10 | dist= 0.607m, dimerr= 46.655mm | error: 34.019pix --> 10.548mm

Calibration params (roll,pitch,yaw,x,y,z): -1.2984,0.0135,-1.6493,0.2423,0.0120,-0.0947

Mean reprojection error across 10 samples

  • Error (pix) = 27.601 pix, stdev = 13.723
  • Error (mm) = 11.876 mm , stdev = 8.338
Image

c51-source avatar Aug 08 '25 04:08 c51-source

你的config的yaml是什么，可以发一下，讨论一下

GISer1909 avatar Aug 09 '25 12:08 GISer1909

你的yaml中都是怎么设置的呀，和avia一样吗

zhangqi-sudo avatar Aug 26 '25 02:08 zhangqi-sudo

我利用速腾16线雷达+d435相机去复现过程中轨迹是正确的，但是点云是白色的，偶尔会有一点点错乱的颜色是什么原因，利用message_filters做的软同步，标定方法是利用cam_lidar_calibration，得到rpyxyz在换算成外参的旋转矩阵，下面是我的标定结果以及运行算法的截图 process[visualise_results-1]: started with pid [15846] process[assess-2]: started with pid [15847] [INFO] [1753099156.564098074]: Importing samples from: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/poses.csv [INFO] [1753099156.564558913]: 10 samples imported [INFO] [1753099157.032299]: Starting visualise_results [INFO] [1753099157.035458]: Opening file at: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/calibration_2025-07-21_19-57-02.csv [INFO] [1753099157.036326]: Using degrees for rotation

Total number of samples: 44 roll | Num samples after filtering: 43 pitch | Num samples after filtering: 42 yaw | Num samples after filtering: 44 x | Num samples after filtering: 41 y | Num samples after filtering: 44 z | Num samples after filtering: 44

---- Calculating average reprojection error on 10 samples ---- 1/ 10 | dist= 0.699m, dimerr= 62.015mm | error: 14.597pix --> 5.516mm 2/ 10 | dist= 0.745m, dimerr= 64.334mm | error: 25.940pix --> 10.351mm 3/ 10 | dist= 0.740m, dimerr= 59.449mm | error: 34.377pix --> 14.200mm 4/ 10 | dist= 0.591m, dimerr= 41.913mm | error: 17.080pix --> 6.921mm 5/ 10 | dist= 1.111m, dimerr= 49.753mm | error: 50.072pix --> 31.171mm 6/ 10 | dist= 0.710m, dimerr= 37.961mm | error: 11.019pix --> 4.552mm 7/ 10 | dist= 0.687m, dimerr= 53.206mm | error: 12.545pix --> 4.547mm 8/ 10 | dist= 0.810m, dimerr= 32.256mm | error: 45.064pix --> 20.254mm 9/ 10 | dist= 0.663m, dimerr= 32.691mm | error: 31.299pix --> 10.704mm 10/ 10 | dist= 0.607m, dimerr= 46.655mm | error: 34.019pix --> 10.548mm

Calibration params (roll,pitch,yaw,x,y,z): -1.2984,0.0135,-1.6493,0.2423,0.0120,-0.0947

Mean reprojection error across 10 samples

  • Error (pix) = 27.601 pix, stdev = 13.723
  • Error (mm) = 11.876 mm , stdev = 8.338
Image

你好，我也在使用robosense雷达和intel realsense系列相机复现，方便加个qq交流一下吗？

wzhaofeng avatar Sep 18 '25 05:09 wzhaofeng

可以的

---原始邮件--- 发件人: @.> 发送时间: 2025年9月18日(周四) 中午1:18 收件人: @.>; 抄送: @.@.>; 主题: Re: [hku-mars/FAST-LIVO2] 速腾16线雷达+d435相机可以复现嘛？ (Issue #349)

wzhaofeng left a comment (hku-mars/FAST-LIVO2#349)

我利用速腾16线雷达+d435相机去复现过程中轨迹是正确的，但是点云是白色的，偶尔会有一点点错乱的颜色是什么原因，利用message_filters做的软同步，标定方法是利用cam_lidar_calibration，得到rpyxyz在换算成外参的旋转矩阵，下面是我的标定结果以及运行算法的截图 process[visualise_results-1]: started with pid [15846] process[assess-2]: started with pid [15847] [INFO] [1753099156.564098074]: Importing samples from: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/poses.csv [INFO] [1753099156.564558913]: 10 samples imported [INFO] [1753099157.032299]: Starting visualise_results [INFO] [1753099157.035458]: Opening file at: /home/bobi/lidar_camera02/src/cam_lidar_calibration-master/data/2025-07-21_18-56-49/calibration_2025-07-21_19-57-02.csv [INFO] [1753099157.036326]: Using degrees for rotation

Total number of samples: 44 roll | Num samples after filtering: 43 pitch | Num samples after filtering: 42 yaw | Num samples after filtering: 44 x | Num samples after filtering: 41 y | Num samples after filtering: 44 z | Num samples after filtering: 44

---- Calculating average reprojection error on 10 samples ---- 1/ 10 | dist= 0.699m, dimerr= 62.015mm | error: 14.597pix --> 5.516mm 2/ 10 | dist= 0.745m, dimerr= 64.334mm | error: 25.940pix --> 10.351mm 3/ 10 | dist= 0.740m, dimerr= 59.449mm | error: 34.377pix --> 14.200mm 4/ 10 | dist= 0.591m, dimerr= 41.913mm | error: 17.080pix --> 6.921mm 5/ 10 | dist= 1.111m, dimerr= 49.753mm | error: 50.072pix --> 31.171mm 6/ 10 | dist= 0.710m, dimerr= 37.961mm | error: 11.019pix --> 4.552mm 7/ 10 | dist= 0.687m, dimerr= 53.206mm | error: 12.545pix --> 4.547mm 8/ 10 | dist= 0.810m, dimerr= 32.256mm | error: 45.064pix --> 20.254mm 9/ 10 | dist= 0.663m, dimerr= 32.691mm | error: 31.299pix --> 10.704mm 10/ 10 | dist= 0.607m, dimerr= 46.655mm | error: 34.019pix --> 10.548mm

Calibration params (roll,pitch,yaw,x,y,z): -1.2984,0.0135,-1.6493,0.2423,0.0120,-0.0947

Mean reprojection error across 10 samples

Error (pix) = 27.601 pix, stdev = 13.723

Error (mm) = 11.876 mm , stdev = 8.338

你好，我也在使用robosense雷达和intel realsense系列相机复现，方便加个qq交流一下吗？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

c51-source avatar Sep 18 '25 05:09 c51-source

可以的 … 太好了，我的qq1526335437，感谢

wzhaofeng avatar Sep 18 '25 06:09 wzhaofeng