This project leverages PyTorch's MiDaS model and Open3D's point cloud implementation to attempt to create an orthogonal 3D mapping of a scene. The goal of this project is to provide a real-time, accurate, and intuitive representation of the spatial relationships between objects in a given scene.
This project leverages PyTorch's MiDaS model and Open3D's point cloud implementation to attempt an orthogonal 3D mapping of a scene.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
The project build is tested on Python 3.9 and 3.10. For Mac and Linux environments, it is recommended that you manage your versions with pyenv, or (for the general case) use a virtual env to avoid clashing dependencies with Torch packages.
A step by step series of examples that tell you how to get a development env running.
First, you need to install the Python requirements.
pip install -r requirements.txtTo run the model on an image, replace FILENAME with the file name (with path and extension) of the image to be used as input. For the --level configuration option, input an integer 1-3, 1 being the lowest accuracy with highest inference speed and 3 being highest accuracy with slowest inference speed.
python3 main.py --level [1|2|3] --image FILENAME- Hat tip to anyone whose code was used
- Inspiration
- References