Omega-Prime-Visibility

This packages uses VisiLibity1 through visilibity (LGPL) to comptue the visibility of MovingObjects of omega-prime-Recordings. For every object it computes the visibilty as a float (1.0 for fully visible and 0.0 for not visible) (assuming 2D-brids-eye-view geometries) from the point of view of one object (assumed as a point in the center of the object). In addition the objects occluding the view (occluders) are computed.

This package also provides the computation of visibility as an omega-prime Metric: from omega_prime_visibility import visibility.

The example file highway_merge.mcap is taken from omega-prime and derived from esmini and is under MPL-2.0 license.

License

This python package is distributed under MIT license but the linked libraries visilibity and VisiLibity1 are under the GNU Lesser General Public License (LGPL).

Requirements

You need to have installed swig, boost and cython to be able to install visilibity

On windows you need to additonally install Buildtools:

1. Download Buildtools für Visual Studio 2022
2. Select C++ Tools for Linux Development and install

Installation

pip install omega-prime-visibility

Usage

See ./tutorial.ipynb for detailed instructions.

from omega_prime_visibility import get_visibility_df
import omega_prime
import shapely

r = omega_prime.Recording.from_file('highway_merge.mcap', compute_polygons=True)

obstruction_poly = shapely.Polygon([
    [-220,10],
    [-120,10],
    [-120,-0],
    [-220,-0],
])
df = get_visibility_df(r._df, ego_idx=0, static_occluder_polys=[obstruction_poly])

returns

frame	idx	occluder_idxs	static_occluder_idxs	visibility
0	1	[]	[0]	0.42
0	2	[]	[0]	0
0	3	[]	[0]	0
0	4	[]	[0]	0
0	5	[3]	[0]	0
1	1	[]	[0]	0.53
1	2	[]	[0]	0
1	3	[]	[0]	0
1	4	[]	[0]	0
1	5	[3]	[0]	0
2	1	[]	[0]	0.63
2	2	[]	[0]	0
2	3	[]	[0]	0
2	4	[]	[0]	0
2	5	[3]	[0]	0
3	1	[]	[0]	0.74
...	...	...	...	...
432	3	[]	[]	1
432	4	[1 2 5]	[]	0
432	5	[1 2]	[]	0

Algorithm

For every point in time compute a visibility for every object from the perspective of an ego vehicle based on 2D-brids-eye-view geometry:

1. Assume point of obersvation to be the centerpoint of the polygon of the ego vehicle
2. Take polygons of all other objects and create visibility polygon using VisiLibity1 through PyVisiLibity.
3. Create shapely.STRTree from the same polygons
4. For every object we want the visibilty for:
   1. Compute the angle range of the object from the perspective of the ego: e.g., The vectors from the perspective point to all points of the polygon have a angle between 90° and 100°. This give a maximum viewable angle-range of 10°.
   2. Compute the visible sections of the polygon by intersecting the polygon with the visibility polygon (speed up by using the STRTree)
   3. Compute the angle-range of all visibile segments of the polygon: e.g., [90°,92°] and [98°-100°] are the resulting ranges. This gives a total viewable angle range of 2°+2°=4°
   4. Divide the viewable angle range by the maximum angle range resulting in the visibility between 0 and 1: e.g., 4°/10°=0.4.
   5. Create the convex hull of the polygon and the persepctive point and find all objects that are interesecting this hull. These are classified as the occluders.

Notice

Important

The project is open-sourced and maintained by the Institute for Automotive Engineering (ika) at RWTH Aachen University. We cover a wide variety of research topics within our Vehicle Intelligence & Automated Driving domain. If you would like to learn more about how we can support your automated driving or robotics efforts, feel free to reach out to us! :email: opensource@ika.rwth-aachen.de