maxibooksiyi commented 4 months ago

我用FAST-LIVO跑MARS Dataset时无人机一起飞终端就会挂掉，换了好几个序列都是这样，包括AMtown AMvalley HKisland序列，无人机刚开始还在地面上放置时都是可以正常运行一段时间的，无人机一起飞到半空中，终端就会报红挂掉，或则我改为bag包播放到无人机飞到空中时再启动FAST-LIVO，终端基本上立马就报红挂掉，相机内参都已经改为MARS Dataset对应的相机内参了。想知道挂掉的原因是什么，以及该如何解决... 下面是挂掉时的一些终端截图

这是我的avia_resize.yaml的内容
`feature_extract_enable : 0 point_filter_num : 2 max_iteration : 10 dense_map_enable : 1 filter_size_surf : 0.15 filter_size_map : 0.3 cube_side_length : 20 debug : 0 grid_size : 40 patch_size : 8 img_enable : 1 lidar_enable : 1 outlier_threshold : 300 # 78 100 156 ncc_en: false ncc_thre: 0 img_point_cov : 100 # 1000 laser_point_cov : 0.001 # 0.001 cam_fx: 1453.72 cam_fy: 1453.28 cam_cx: 1172.18 cam_cy: 1041.78

common: lid_topic: "/livox/lidar" imu_topic: "/livox/imu"

preprocess: lidar_type: 1 # 1:Livox Avia LiDAR 2:VELO16 3:OUST64 4:XT32 scan_line: 6 # 16 64 32 blind: 5 # blind x m disable

mapping: acc_cov_scale: 100 gyr_cov_scale: 10000 fov_degree: 90 # 360 extrinsic_T: [ 0.04165, 0.02326, -0.0284 ] extrinsic_R: [ 1, 0, 0, 0, 1, 0, 0, 0, 1]

pcd_save: pcd_save_en: true

camera:

img_topic: /usb_cam/image_raw

# img_topic:  /camera/image_color
img_topic: /left_camera/image
#xiyuan
Rcl: [0.00162756,-0.999991,0.00390957,
     -0.0126748,-0.00392989,-0.999912,
      0.999918,0.00157786,-0.012681]
Pcl: [0.0409257, 0.0318424, -0.0927219]
`

xuankuzcr commented 4 months ago

我刚才测了一下感觉没什么太大问题，截图结果如下。另外我一会儿单独为MARS-LVIG dataset创建一个配置文件然后push上去吧，方便大家复现，之后你可以再测试下哈。但最好你能用gdb检查下到底问题出在哪行#53，我比较担心是你的环境配置问题。 rviz_screenshot_2024_07_19-03_33_40 Screenshot from 2024-07-19 12-14-12

maxibooksiyi commented 4 months ago

非常感谢您的回答，很有帮助，我用gdb调试后打印信息如下

[ VIO ]: Add 0 3D points.
[ VIO ]: time: addFromSparseMap: 0.000011 addSparseMap: 0.000055 ComputeJ: 0.000004 addObservation: 0.000002 total time: 0.000072 ave_total: 0.000072.
[ LIO ]: Raw feature num: 11893 downsamp num 10911 Map num: 10964.
[ LIO ]: Using multi-processor, used core number: 4.
[New Thread 0x7fff923f8700 (LWP 18464)]
[New Thread 0x7fff91bf7700 (LWP 18465)]
[New Thread 0x7fff913f6700 (LWP 18466)]
[ LIO ]: time: fov_check: 0.000000 fov_check and readd: 0.002638 match: 0.000000 solve: 0.000000  ICP: 0.001149  map incre: 0.031386 total: 0.035173 icp: 0.000780 construct H: 0.000000.
[ INFO ]: get point cloud at time: 1658137064.324880.
[ VIO ]: Raw feature num: 11893.
[ VIO ]: Add 1761 3D points.
[ VIO ]: time: addFromSparseMap: 0.000001 addSparseMap: 0.006574 ComputeJ: 0.000000 addObservation: 0.000000 total time: 0.006576 ave_total: 0.006576.
[ INFO ]: get img at time: 1658137064.440385.
[ LIO ]: Raw feature num: 11898 downsamp num 10905 Map num: 17091.
[ LIO ]: Using multi-processor, used core number: 4.
[ LIO ]: time: fov_check: 0.000000 fov_check and readd: 0.002378 match: 0.000000 solve: 0.000000  ICP: 0.002737  map incre: 0.020595 total: 0.030441 icp: 0.001689 construct H: 0.000000.
[ INFO ]: get img at time: 1658137064.540787.
[ INFO ]: get point cloud at time: 1658137064.424726.
[ VIO ]: Raw feature num: 11898.

Thread 1 "fastlivo_mappin" received signal SIGSEGV, Segmentation fault.
0x00007ffff32a6088 in lidar_selection::LidarSelector::addFromSparseMap (
    this=this@entry=0x55555c87d3c0, img=..., pg=...)
    at /home/maxi/FASTLIVO_ws/src/FAST-LIVO/src/lidar_selection.cpp:376
376         float it[height*width] = {0.0};
(gdb)

然后参考您给的 https://github.com/hku-mars/FAST-LIVO/issues/53 ，我把lidar_selection.cpp的376行的float it[height width] = {0.0};改为了std::vector it(heightwidth, 0); 再重新编译后运行，便可以基于MARS-LVIG dataset跑起来了。
但是运行不了多长时间（不到一分钟），终端就会报红挂掉，从gnome-system-monitor里面可以看到，FAST-LIVO运行起来后，内存占用一直在累积上升，到达百分之百之后程序终端就会卡住报红挂掉。我是普通笔记本上跑的，内存是16G。似乎和这个 https://github.com/hku-mars/FAST-LIVO/issues/105 类似是么。但在跑FAST-LIVO-Datasets里的hku1.bag是可以运行相对长点的时间而且跑完的。

%I_ Q}QLL$8UA8JHM6BN3Z1

xuankuzcr commented 4 months ago

相同的地图分辨率，大尺度场景内存涨的会越快，因为每帧的raw points会有特别多的点被用来更新地图。可以试着调低filter_size_map，关掉rviz，或增加swap空间。另外我已经上传了MARS-LVIG的参数和launch文件，并为了适配做了一些的代码修改。如果想要速度更快，内存占用更低的版本，可以follow下FAST-LIVO2。

hku-mars / FAST-LIVO

跑MARS Dataset时无人机一起飞终端就会报红挂掉 #108

img_topic: /usb_cam/image_raw