pcl_catkin && catkin build dcl_fast_lio error

[Lee-JaeWon]

I tried running dcl-slam on ROS noetic, but it doesn't work.

I cloned and built pcl_catkin github in the src folder because I think I need a pcl-catkin package when I build.

However, when I follow your README step by step, an error occurs in the process of catkin build dcl_fast_lio.

Do you know the solution?

I am confident that I have satisfied both the version and the dependency.

Specifically, I think I need to know what's different between your environment and mine. Especially for the pcl_catkin package.

To sum up, I only added pcl_catkin to your README method. If I don't add pcl_catkin, it says pcl_catkin is required, and if I add it, another error occurs.

This relates to issue #14 .

I'd appreciate it if you could answer.

[zhongshp]

Hi~ Lee-JaeWon
Thank you for pointing it out. The current DCL-SLAM code is only compatible with Ubuntu 18.04 Melodic. However, the author also found a compilation problem with the pcl library in Ubuntu 20.04 noetic, which appears to be a conflict between version 1.8 of Pcl-Catkin and ubuntu 20.04 PCL-1.10 (The code for ubuntu 20.04 is coming soon). My suggestion is to skip pcl-catkin and use pcl-1.10, which comes with ubuntu 20.04. Introduction of pcl-1.10 in cmakelists:

find_package(PCL REQUIRED)
include_directories(
${PCL_INCLUDE_DIRS}
)
link_directories(
${PCL_LIBRARY_DIRS}
)

And when compiling the Node executable:
target_link_libraries(Node ${PCL_LIBRARIES})

[heisetiantang]

作者您好，很荣幸你能回答之前的问题，关于DCL的复现，还有些问题想向您请教，通过阅读您的论文和代码，一直没有完全弄懂，在三个机器人之间，程序需要传递那几个话题，才能够使得实际机器人，像数据集跑得那样，能够多机器人协同定位建图，或者说通过什么能够证实多个机器人之间成功进行协同，仅仅通过观察地图的变换么，还是在机器人进行回环后，可以接收到其他机器人的位姿和地图，对当前机器人进行补全呢。

[heisetiantang]

@zhongshp @Lee-JaeWon

[Lee-JaeWon]

@zhongshp
Thank you for your answering.
But ROS melodic also has problems with pcl_catkin.
I installed pcl 1.8.1(in ubuntu) and build fast_lio by cloning the appropriate pcl_catkin in github, but I get an error.
It seems necessary to organize this dependency problem, and I would appreciate it if you could organize it.

LIO-SAM works well. Only fast-lio seems to need a solution.

All were purely produced and executed in docker environments.