There are many algoritmhs which are used for plane segmentation, especially based on RANSAC technique. Despite the high accuracy given by RANSAC, the method has its own limitations.
There is class of plane segmentation algorithms which are tuned to work with data obtained from RGB-D Camera. The existing implemtation of most of them is written in C++, so it may be hard to build and configure each of the algorithms.
plane-seg is meant to be an easy-to-use wrapper for running and configuring those algorithms. To run algorithm we use Docker images from 3D-plane-segmentation. More than that we provide interface for evaluating the results of segmentation based on EVOPS package.
All Python dependencies are specified in python/requirements.txt and can be installed with pip or another package manager.
If you want to run any of the wrapped algorithms, you need to have applicable Docker image installed on your system:
| Algorithm | Docker image |
|---|---|
| DDPFF | image |
| peac | image |
| CAPE | image |
You may find examples of running the algorithm in python/examples as well as evaluating metrics on segmentation result.
Apache License, Version 2.0