How to efficiently build a models database of home furnishings
Hi all,
I'd like to train an object detector to recognize home furnishings (e.g., tables, chairs, sofas, etc...). To do so, I considered the idea of building several indoor environments in gazebo to generate the training set. So, I'm faced with the problem of retrieving a considerable number of 3D models for each object category and import them into gazebo.
My first attempt has been to download collada models from the 3D Warehouse and import them in the Model Editor to create an sdf model. The problem with this procedure is that it's not always working (e.g., missing parts of the model and/or missing textures). My guess is that the available collada files are buggy and need some editing.
So, I'm posting this question because I'd be very glad if someone could point me out an efficient way to solve this problem or at least give some suggestions on how to proceed!
Thanks.
Asked by federico.nardi on 2017-11-24 04:28:48 UTC
Comments