making it look nicer

This commit is contained in:
Sabrina
2020-04-30 21:55:11 -04:00
committed by GitHub
parent 74d38c1cd9
commit c0d753a7e5

View File

@@ -91,49 +91,76 @@ Install the following dependencies
2. `pip install numpy matplotlib open3d-python progressbar pyquaternion transforms3d scipy scikit-image networkx psutil`
3. Install required dependencies including ray_tracing, pybind11, pytest, and Catch2
`git clone https://github.com/catchorg/Catch2.git`
`mkdir Catch2/build`
`cd Catch2/build`
`cmake ..`
`make -j8`
`sudo make install`
`pip install pytest`
`git clone https://github.com/pybind/pybind11.git`
`mkdir pybind11/build`
`cd pybind11/build`
`cmake ..`
`make -j8`
`sudo make install`
`git clone https://github.com/acschaefer/ray_tracing.git`
`mkdir ray_tracing/build && cd ray_tracing/build`
`cmake ..`
`make -j8`
`sudo make install`
4. Download raw data from [NCLT](http://robots.engin.umich.edu/nclt/) to run code from scratch OR use our [prebuilt maps](https://drive.google.com/drive/folders/1cFf0q76xyul4nbShm-GwDNxFwYh1Bkzx?usp=sharing)
4.1 The raw data can be downloaded from the website directly or using the downloader.py. If you are using the prebuilt maps, skip this step.
`cd src/dataset/`
`python download.py --date="2012-01-15" --vel --gt --gt_cov --sensor`
`cd nclt/sensors/<date>`
`tar xzf <sensors_file>`
`cd ../velodyne`
`tar xzf <velodyne_file>`
Repeat this for any interested datasets. We have maps made for 2012-01-15, 2012-01-08, 2013-04-05, and 2013-01-10.
5. In ncltpoles.py we need to make a couple changes
`cd src/polex/poles/`
Open and edit ncltpoles.py with your favorite editor.
There are two # TODO comments. In the first one change 2020-04 to <year>-<month>. In the second one change the session to your desired session date. If you would like to run the exact experiment we did, do not change this date.
6. In pynclt.py we need to make a couple changes
`cd src/polex/poles`
Open and edit pynclt.py with your favorite editor.
There are two # TODO comments. In the first one, change paths to the appropraite directories on your machine. In the second one, comment our the sessions you are not using.
7. Run it. WARNING this can take anywhere from 1 - 10 hours for 4 datasets.
`cd src/polex/poles`
`python2 ncltpoles.py`