Changes to the original loam_velodyne code

[YoshuaNava]

Hey,
I'm Yoshua, robotics engineer working in Zürich, Switzerland but originally from Venezuela. Nice to meet you!

I'm interested in the work you're doing. I've been doing research on LiDAR SLAM and loop closure detection, with the purpose of further increasing the accuracy of LOAM, obtaining both globally consistent and locally accurate maps. As you, I have observed that loam_velodyne from laboshinl doesn't give highly accurate results as the ones claimed by Zhang et al in their papers.

I wanted to ask you, first of all, what kind of changes you did to the LOAM code, and if you took any of the existing LOAM forks as a starting point (for example: laboshinl's fork, lego-loam, the original version from zhang, martin-velas, etc). I would be interested in collaborating to make the open source version of LOAM as accurate as possible.

Thank you.

[claydergc]

Hello Yoshua, Nice to meet you, too. I am from Peru and I am doing my masters at Universidad de Chile. What I did was taking laboshinl's version with the last commit done by Adi Shavit, which decouples ROS from the core LOAM algorithms. After that I just modified ScanRegistration.cpp, MultiScanRegistration.cpp, LaserOdometry.cpp and TransformMaintenance.cpp to get transform matrices based on Eigen library instead of ROS odometry topics. Finally, I added a main.cpp where I create instances of MultiScanRegistration, LaserOdometry, LaserMapping and TransformMaintenance and call the functions I created before to get the /integrated_to_init odometry topic as a transform matrix. I use that matrix to compare the odometry with the ground-truth of the KITTI Dataset, but I am not getting good results. Your collaboration would be great given that I am using this algorithm for my master thesis. Let me know If you want to make some change to my code or if you want to create a new one. Thanks. El vie., 31 ago. 2018 a las 13:59, Yoshua Nava (<notifications@github.com>) escribió:

…

Hey, I'm Yoshua, robotics engineer working in Zürich, Switzerland but originally from Venezuela. Nice to meet you! I'm interested in the work you're doing. I've been doing research on LiDAR SLAM and loop closure detection, with the purpose of further increasing the accuracy of LOAM, obtaining both globally consistent and locally accurate maps. As you, I have observed that loam_velodyne from laboshinl doesn't give highly accurate results as the ones claimed by Zhang et al in their papers. I wanted to ask you, first of all, what kind of changes you did to the LOAM code, and if you took any of the existing LOAM forks as a starting point (for example: laboshinl's fork, lego-loam, the original version from zhang, martin-velas, etc). I would be interested in collaborating to make the open source version of LOAM as accurate as possible. Thank you. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKjpF0-Z2TGfjJbD8x0Bzg3zEUVRt8bmks5uWYekgaJpZM4WVtcQ> .

-- Clayder Gonzalez Cadenillas

[jianrui1]

@claydergc I have done as you code. But a problem appeared.As follow:
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_M_construct null not valid
why?can you tell me the reason and help me out?
Thank you so much