Hello I am new here and work in a thesis for the first time with a LIDAR file and thus with CloudCompare. I have a LAS-file which reflects a test drive on a LOK. The goal is to automatically detect objects in this file, such as signals or Switches and stuff like this. Only I have several problems there.
1st problem was that the file was 100 GB in size, which is why I have now divided these into "tiles" using LAStile to work well with it. I think that's a solution that will be enough for this time.
My bigger second problem i have right now is, the problem that there is an LOK artifact throughout the ride, which makes evaluation difficult, so I now have to find a solution to how to remove that "Fault".
I have to fixed description of the fault with maybe the coordinates. But the problem is, i miss the trajectory in the file and when i view the tiles in CloudCompare. Is there a way to define this trajectory by myself?
Can I set the center of the traveled route in CloudCompare and then always filter out the same interference using fixed distances and Fixed points in distance to the trajectory?
I hope I could describe my problem understandable and it is not too banal, have a great day!(:
Object Detection in a LIDAR File of a Test drive with a LIDAR system on a train through a track
Re: Object Detection in a LIDAR File of a Test drive with a LIDAR system on a train through a track
Sorry, I don't understand your problem well enough to be able to help.
LAS files don't contain the trajectory of the vessel. But sometimes you have the GPS time and the scan angle rank, etc. that may help 'guessing' it.
But if you have this trajectory as a polyline for instance in CC, you won't be able to compute the distance to it ;) (it's better to have it as a - dense - point cloud).
P.S.: what is LOK?
LAS files don't contain the trajectory of the vessel. But sometimes you have the GPS time and the scan angle rank, etc. that may help 'guessing' it.
But if you have this trajectory as a polyline for instance in CC, you won't be able to compute the distance to it ;) (it's better to have it as a - dense - point cloud).
P.S.: what is LOK?
Daniel, CloudCompare admin
Re: Object Detection in a LIDAR File of a Test drive with a LIDAR system on a train through a track
Thanks for your fast Feedback,
working with point clouds is realy new for my, because of that i maybe asked stuped things.
Yes i checked the LAS-filetype and i realised there is no information of a trajectory, so i checked again the data from the test drive and i now have a .kml-data with the informations of the trajectory.
I now have to tasks i would like to manage.
1. I would like to automaticly delete a "reflection" of the Train (where the Laserscanner was installed). So my idea is to use the trajectory as the relation to delete all the point in a area above the trajectory, because there could not be any points i have to find.
2. I would like to start a test with automatic object recognition with a kind of classification. The objects i would like to find are a) usefull objects and b) unusable objects (like trees). I startet to sort the objects by characteristics.
Objects i would like to find are for example Light signals of the railway.
I used LAStools (LAStiles) to have a smaler, usefull area with this objects.
I also used the CSF plugin to have groundet and off-groundet points.
Now i would like to classify some points in the off-groundet points. Is there a way to search for a object with typicall size x/y/z, and could u "scan" one signal, save the size and search for this object?
Thanks for your help!
PS: "Lok" is the german word for a "railway engine", sorry!
working with point clouds is realy new for my, because of that i maybe asked stuped things.
Yes i checked the LAS-filetype and i realised there is no information of a trajectory, so i checked again the data from the test drive and i now have a .kml-data with the informations of the trajectory.
I now have to tasks i would like to manage.
1. I would like to automaticly delete a "reflection" of the Train (where the Laserscanner was installed). So my idea is to use the trajectory as the relation to delete all the point in a area above the trajectory, because there could not be any points i have to find.
2. I would like to start a test with automatic object recognition with a kind of classification. The objects i would like to find are a) usefull objects and b) unusable objects (like trees). I startet to sort the objects by characteristics.
Objects i would like to find are for example Light signals of the railway.
I used LAStools (LAStiles) to have a smaler, usefull area with this objects.
I also used the CSF plugin to have groundet and off-groundet points.
Now i would like to classify some points in the off-groundet points. Is there a way to search for a object with typicall size x/y/z, and could u "scan" one signal, save the size and search for this object?
Thanks for your help!
PS: "Lok" is the german word for a "railway engine", sorry!