Hello!
I have a question regarding classifying objects in a point cloud model if the model has several different off-ground objects also- like trees, cars, traffic signs. What would be the best way to create different classes for each one of them, I have classified the ground and all the other objects are just one class- off-ground. Also is it possible to segment objects in those classes? Like there is a class of cars and each car is segmented separately as car1, car2,... etc?
beginner question here
Re: beginner question here
Are you referring to the CANUPO plugin maybe?
If yes, then you need to create several binary classifiers to segment the off ground points, typically between cars and 'not cars', and then another one to differentiate between 'car type 1' and 'car type different from 1', etc.
That can become tedious, and also, I am not sure CANUPO is well suited to get to this level of detail.
If yes, then you need to create several binary classifiers to segment the off ground points, typically between cars and 'not cars', and then another one to differentiate between 'car type 1' and 'car type different from 1', etc.
That can become tedious, and also, I am not sure CANUPO is well suited to get to this level of detail.
Daniel, CloudCompare admin
Re: beginner question here
Hi. Can I segment (with scissors tool) my point cloud for multiple samples? i.e. I want the classifier to have more examples to classify afterwards.
I mean I want to tell the classifier to use 2 sample (polygons cut) for class 1 and 2 samples for class 2 to make classifier more exact. I realized I can't use some like CTRL+<click> or others to do that. Anyone can help?
I mean I want to tell the classifier to use 2 sample (polygons cut) for class 1 and 2 samples for class 2 to make classifier more exact. I realized I can't use some like CTRL+<click> or others to do that. Anyone can help?
Re: beginner question here
You can segment the cloud multiple times, and then merge the parts back together (with 'Edit > Merge')
Daniel, CloudCompare admin