Skip to content

petervolvowinz/Waymo-visu

Repository files navigation

Simple Waymo Open Motion Dataset Reader

WAYMO motion data

This is a simple file reader and visualizer for the waymo open motion dataset. This is originally cloned from Verses which in turn cloned from WAYMO repo.

https://github.com/waymo-research/waymo-open-dataset

We have added animation of scenarios from the waymo data set focusing on the ego vehicle and the suggested track prediction array. The animation is saved in a mp4 video file for playback - in the data folder. Pedestrians are drawn as red circles, green bounding boxes are bicylists. The data is projected in 2D.

Installation

For dependencies, you need protobuf, tensorflow, matplotlib, numpy:

pip install protobuf, tensorflow, matplotlib, numpy

For matplotlib you need to choose correct working env.

matplotlib.use('MacOSX')

Then install the repo:

pip install -e .

Optionally regenerate the *_pb2.py files from the protos:

sh generate_proto.sh

Usage

Under the data folder there you need to download a sample scenario record - the code refers to this file(training.tfrecord-00014-of-01000). Please look here to access (github wont allow larger files than 100 MB): https://waymo.com/open/download/. Accessing the data set directly can be also be done:

FILENAME2 = 'gs://waymo_open_dataset_motion_v_1_2_1/uncompressed/scenario/training/training.<fileextension>'

In the data folder the animation is saved with name training.tfrecord-00014-of-01000_all.mp4 All refers to the fact that all objects - vehicles, pedestrians, cyclists - are animated. The flag in scenario_animation.py is used to toggle between all and limited object list.

use_all_data_objects = True

Above is a way to access the dataset directly.

Examples of animation and dataset traversal.

    parse_motion_dataset.py
    count_frames.py
    animationtest.py
    zoomeffects.py

Parsing and visualization scripts:

count_frames.py             # traverse and count motion data frames per file
motion-scene-withzoom.py    # traverse and plot a zoomable subset
parse_motion_dataset.py     # parse and count pedestrians per dataset
plot_vehicle.py             # parse and animate the scenario with field of view from ego vehicle

License

This code is released under the Apache License. This projects incorporate some parts of the Waymo Open Dataset code (the files simple_waymo_open_dataset_reader/*.proto) and is licensed to you under their original license terms. See LICENSE file for details.

About

Way motion data set visualization

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published