sensors

Article
Semantic Evidential Grid Mapping Using Monocular and
Stereo Cameras †
Sven Richter 1, * , Yiqun Wang 1 , Johannes Beck 2 , Sascha Wirges 1 and Christoph Stiller 1

1 Institute of Measurement and Control Systems, Karlsruhe Institute of Technology (KIT), Engler-Bunte-Ring 21,
76131 Karlsruhe, Germany; [email protected] (Y.W.); [email protected] (S.W.);
[email protected] (C.S.)
2 Atlatec GmbH, Haid-und-Neu-Straße 7, 76131 Karlsruhe, Germany; [email protected]
* Correspondence: [email protected]
† This paper is an extended version of our paper published in Richter, S.; Beck, J.; Wirges, S.; Stiller, C. Semantic
Evidential Grid Mapping based on Stereo Vision. In Proceedings of the 2020 IEEE International Conference
on Multisensor Fusion and Integration for Intelligent Systems (MFI), Karlsruhe, Germany, 14–16 September
2020; pp. 179–184, doi:10.1109/MFI49285.2020.9235217.

Abstract: Accurately estimating the current state of local traffic scenes is one of the key problems
in the development of software components for automated vehicles. In addition to details on free
space and drivability, static and dynamic traffic participants and information on the semantics may
also be included in the desired representation. Multi-layer grid maps allow the inclusion of all of
this information in a common representation. However, most existing grid mapping approaches
only process range sensor measurements such as Lidar and Radar and solely model occupancy
without semantic states. In order to add sensor redundancy and diversity, it is desired to add vision-



based sensor setups in a common grid map representation. In this work, we present a semantic


evidential grid mapping pipeline, including estimates for eight semantic classes, that is designed
Citation: Richter, S.; Wang, Y.;
for straightforward fusion with range sensor data. Unlike other publications, our representation
Beck, J.; Wirges, S.; Stiller, C. Semantic
explicitly models uncertainties in the evidential model. We present results of our grid mapping
Evidential Grid Mapping Using
pipeline based on a monocular vision setup and a stereo vision setup. Our mapping results are
Monocular and Stereo Cameras.
Sensors 2021, 21, 3380. https://
accurate and dense mapping due to the incorporation of a disparity- or depth-based ground surface
doi.org/10.3390/s21103380 estimation in the inverse perspective mapping. We conclude this paper by providing a detailed
quantitative evaluation based on real traffic scenarios in the KITTI odometry benchmark dataset and
Academic Editors: Sukhan Lee, Uwe demonstrating the advantages compared to other semantic grid mapping approaches.
D. Hanebeck and Florian Pfaff
Keywords: autonomous driving; environment perception; grid mapping; stereo vision; monocular
Received: 4 April 2021 vision
Accepted: 7 May 2021
Published: 12 May 2021

Publisher’s Note: MDPI stays neutral 1. Introduction

with regard to jurisdictional claims in Environment perception modules in automated driving aim at solving a wide range
published maps and institutional affil-
of tasks. One of these is the robust and accurate detection and state estimation of other
iations.
traffic participants in areas that are observable by on-board sensors. For risk assessment of
the current scene, information about unobservable areas is also important. Furthermore,
drivable areas must be perceived in order to navigate the automated vehicle safely. To
reduce computational power, a common framework for solving all of these tasks is de-
Copyright: © 2021 by the authors. sirable. Additionally, it is preferable to use multiple heterogeneous sensors to increase
Licensee MDPI, Basel, Switzerland. the robustness of the whole system. In the literature, occupancy grid maps are frequently
This article is an open access article considered, as they enable the detection of other traffic participants while additionally
distributed under the terms and modeling occlusions due to their dense grid structure. Most of the presented methods only
conditions of the Creative Commons
include the processing of range sensor measurements such as Lidar and Radar and solely
Attribution (CC BY) license (https://
model occupancy without semantic states. Cameras have received less attention in the
creativecommons.org/licenses/by/
past couple of years as Lidar sensors have become more and more affordable. However,
4.0/).

Sensors 2021, 21, 3380. https://doi.org/10.3390/s21103380 https://www.mdpi.com/journal/sensors

Sensors 2021, 21, 3380 2 of 16

compared to cameras, Lidar sensors are still more expensive. Furthermore, cameras are
superior when it comes to understanding semantic details in the environment. In [1], we
presented a semantic evidential fusion approach for multi-layer grid maps by introducing
a refined set of hypotheses that allows the joint modeling of occupancy and semantic states
in a common representation. In this work, we use the same evidence theoretical framework
and present two improved sensor models for stereo vision and monocular vision that can
be incorporated in the sensor data fusion presented in [1].
In the remainder of this section, we briefly introduce the terms of the Dempster–Shafer
theory (Section 1.1) relevant to this work. We then review past publications on stereo
vision-based and monocular vision-based grid mapping, monocular depth estimation and
semantic grid mapping in Section 1.2, followed by highlighting our focus for the proposed
methods in Section 1.3. In Section 2, we give an overview of our semantic evidential
models and the multi-layer grid map representations. We further describe our proposed
semantic evidential grid mapping pipelines, depicted in Figure 1, in detail. We evaluate
our processing steps based on challenging real traffic scenarios and compare the results of
both methods in Section 3. Finally, we conclude this paper and give an outlook to future
work in Section 4.

1a Label Histograms 2

Monocular image Monocular depth Ground Surface

Semantic BBA 4

gM : G xy × S → [0, 1]

hM : G M × S → R≥0
Semantic segmentation

Stereo images hM : G xy × S → R≥0

Stereo disparity

1b 3

Figure 1. Overview of the described grid mapping framework. On the front end, both monocular images are processed
to obtain depth maps (1a) or stereo images are used to estimate a disparity map (1b). Both of them are accompanied by a
pixelwise semantic segmentation image. The images are used as input for a label histogram calculation in a setup-dependant
grid in the second step (2). This label histogram is transformed into a cartesian grid (3) and finally transformed into a
semantic evidential grid map (4).

1.1. Dempster–Shafer Theory of Evidence (DST)

The Dempster–Shafer theory of evidence (DST), originally introduced in [2], is an
extension of Bayes theory and provides a framework to model uncertainty and combine
evidence from different sources. For the hypotheses set of interest Ω, called frame of
discernment, the basic belief assignment (BBA)

m : 2Ω → [0, 1] , m(∅) = 0, ∑ m( A ) = 1
A ∈2Ω

assigns belief masses to all possible combinations of evidence. In contrast to probability

measures, the BBA does not define a measure in the measure theoretical sense as it does
not satisfy the additivity property. In consequence, the belief mass assigned to the whole
Sensors 2021, 21, 3380 3 of 16

set Ω models the amount of total ignorance explicitly. Based on a BBA, lower and upper
bounds for the probability mass Pr(·) of a set A ∈ 2Ω can be deduced as

∑ m( B) = bel( A) ≤ Pr( A) ≤ pl(A) = ∑ m( B ),

B⊆ A B∩ A6=∅

where bel(·) and pl(·) are called belief and plausibility, respectively.

1.2. Related Work

Occupancy grid maps, as introduced by Elfes et al. in [3], are often used for dense
scene state estimation as they enable explicit modeling of free space and occlusion. While
cartesian grid maps are especially suitable for fusing measurements over time as, e.g., pre-
sented in [4,5], other coordinate systems are more suitable for modeling sensor charac-
teristics. In [6], Badino et al. compared different tessellations; namely, cartesian, polar
and u-disparity grids for modeling free space in stereo-based occupancy grid maps. Per-
rollaz et al. use a u-disparity grid to estimate a stereo-based occupancy grid map and
further considered different measurement models for assigning pixel measurements to
grid cells in [7]. Danescu et al. used the grid representation from [8] to estimate a dynamic
occupancy grid map with a low-level particle filter in [9]. Yu et al. proposed in [10] to
model free space in a v-disparity grid and occupancy in a u-disparity grid before combining
both in a stereo-based occupancy grid map using an evidence theoretical framework. As
opposed to all previously mentioned works that considered occupancy and free space
only, Giovani et al. added one occupancy refinement value denoting the semantic state to
their grid map representation in [11]. However, as they did not include the semantics in
their evidential hypotheses set, well-established combination rules could not be applied.
Recent work on stereo-based grid mapping has been published by Valente et al. in [12] and
Thomas et al. in [13]. While Valente et al. only modeled occupancy in a u-disparity grid
with a subsequent fusion with Lidar-based occupancy grid maps in the cartesian space,
Thomas et al. incorporated semantic hypotheses in an evidential framework. Focusing
on estimating a road model, however, the hypotheses set they considered is limited to the
static world.
Semantic grid mapping has also been exploited based on measurements from monoc-
ular cameras. Erkent et al. estimated in [14] semantic grid maps by fusing pixel-wise
semantically labeled images with Lidar-based occupancy grid maps in a deep neural net-
work. Lu et al. directly trained a variational encoder–decoder network on monocular
RGB images to obtain a semantic top-view representation in [15]. Both networks result
in a semantic grid map representation containing one class per grid cell, thus discarding
knowledge about the label estimation distribution and uncertainty.
For transforming measurements from the image domain to a top-view representation,
a pixelwise depth estimation is needed. In the last few years, tremendous progress in
monocular depth estimation has been witnessed, especially after the wide deployment and
improvement of deep neural networks. There are three main approaches for monocular
depth estimation with deep neural networks: supervised depth prediction from RGB
images, self-supervised (unsupervised) depth prediction with monocular videos and self-
supervised depth completion.
Nowadays, with the help of convolutional neural networks (CNN), the results, such
as in [16–18], have become superior to previous works in terms of speed and accuracy.
However, the resolution of the monocular depth estimation in those papers is relatively
low. To overcome this predicament, Alhashim et al. present a convolutional neural network
for computing a high-resolution depth map given a single RGB image with the help of
transfer learning [19]. All the above methods attempt to directly predict each pixel’s
depth in an image using models that have been trained offline on a colossal training
dataset with the ground truth images of depth information. While these methods have
enjoyed great success, to date, they have been restricted to scenes where extensive image
collections and their corresponding pixel depths are available. For the case without depth
Sensors 2021, 21, 3380 4 of 16

ground truth dataset, an unsupervised learning framework is presented in [20] for the task
of monocular depth and camera motion estimation from unstructured video sequences.
In [21], the authors generate disparity images from monocular images by training the
network with an image reconstruction loss and stereo images training dataset, exploiting
epipolar geometry constraints. Finally, Qiao et al. tackle the inverse projection problem
in [22] by jointly performing monocular depth estimation and video panoptic segmentation.
With their method, they are able to generate 3D point clouds with instance-level semantic
estimates for each point.

1.3. Goals of This Work

This work aims to provide two accurate and efficient grid mapping frameworks. One
is based on stereo cameras and the other one is based on a monocular camera. In contrast to
many past publications on vision-based grid mapping like [6–8,10,12], we use a wide range
of different semantic classes, which can be provided by vision. Instead of assigning only
one semantic label per grid cell as in [11,14,15], we use the hypotheses set introduced in [1]
to model uncertainty for eight semantic hypotheses in a consistent evidential framework.
In order to achieve a dense and smooth BBA for ground hypotheses, we make use of
encapsulated ground surface estimations to approximate the pixel-to-area correspondence
in the top-view space. The resulting semantic evidential multi-layer grid map can then be
fused with range sensor-based grid maps, as described in [1].

2. Materials and Methods

In this section, we summarize the underlying evidential models, introduce our multi-
layer grid map representations in Section 2.1 and introduce all coordinate systems used
throughout the mapping pipeline in Section 2.2. Then we introduce how to get the input
images for the label histogram calculation in Section 2.3, followed by a detailed description
of the label histogram calculation in the u-disparity and u-depth space in Section 2.4. We
further explain how the label histogram is transformed to a cartesian grid in Section 2.5. We
conclude this section by presenting the calculation of the BBA based on the label histogram
in Section 2.6.

2.1. Semantic Evidential Framework

The frame of discernment

Ω := {c, cy, p, om, nm, s, sw, t}

consists of the hypotheses car (c), cyclist (cy), pedestrian (p), other movable object (om),
non movable object (nm), street (s), sidewalk (sw) and terrain (t). This hypotheses set can
be seen as a refinement of the classical occupancy frame consisting of the two hypotheses
occupied and free by considering the hypotheses sets

O := {{c}, {cy}, { p}, {om}, {nm}} ⊂ 2Ω

and
F := {{s}, {sw}, {t}} ⊂ 2Ω .
This makes it particularly suitable for the fusion of semantic estimates with range
measurements in top-view as outlined in [1]. For the BBA, we consider the hypotheses set
consisting of singletons

S = {{c}, {cy}, { p}, {om}, {nm}, {s}, {sw}, {t}} ⊂ 2Ω

Sensors 2021, 21, 3380 5 of 16

as all hypotheses combinations are either conflicting by definition or not estimated by the
semantic labeling. We define the two-dimensional grid G = P1 × P2 on the rectangular
region of interest R = I1 × I2 ⊂ R2 , where

Pi = { Ii,k , k ∈ {0, . . . , si − 1}},

Ii,k = [oi + k δi , oi + (k + 1) δi ), i ∈ {1, 2}

forms a partition of the interval Ii with equidistant length δi ∈ R, origin oi ∈ R and size
si ∈ N. The BBA m on 2Ω is then represented by the multi-layer grid map

gM : G × S → [0, 1],
(C, ω ) 7→ mC (ω ),

where mC is the corresponding BBA in the grid cell C ∈ G .

2.2. Coordinate Systems

We use four coordinate systems in our processing chain. The first is the image coordi-
nate system with coordinates (u, v) ∈ R2 rectified according to a pinhole camera model.
For mapping stereo vision measurements to the top-view, the u-disparity coordinate system
with coordinates (u, d) ∈ R2 is used as an intermediate representation in order to be able
to model disparity estimation errors explicitly. When depth is estimated directly as in most
of the monocular vision-based methods, a u-depth coordinate system with coordinates
(u, z) ∈ R2 is used. For the final grid representation, a cartesian coordinate system with
coordinates ( x, y) ∈ R2 is used. To indicate the corresponding coordinate system, the con-
sidered region of interest is subscripted as Ruv , Rud , Ruz and R xy , respectively. The same
notation is used for the attached grids Guv , Gud , Guz and G xy . Furthermore, we introduce
the mappings
ud xy
Tuv : Ruv → Rud , Tud : Rud → R xy ,
uz xy
Tuv : Ruv → Ruz , Tuz : Ruz → R xy

for transforming coordinates from one system to another.

2.3. Input Representation

We define a stereo vision measurement
n o
Mstereo = { Pi ∈ Guv , i ∈ {1, . . . , n}}, {fsem , fdisp , fconfdisp } ,

as a tuple of a set of pixels Pi and the three images

fsem : Guv → S , fdisp : Guv → R, fconfdisp : Guv → [0, 1]

which is the pixel-wise semantic labeling image fsem , the disparity image fdisp and disparity
confidence image fconfdisp .
In the case of measurements stemming from a monocular camera, the disparity image
fdisp is replaced by the depth image fdepth : Guv → R:
n o
Mmono = { Pi ∈ Guv , i ∈ {1, . . . , n}}, {fsem , fdepth , fconfdepth }

Note that the confidence images may be set to one for all pixels in case the disparity or
depth estimation does not output one. In this case, every pixel is attached the same weight
in the subsequent grid mapping pipeline. Figure 2 shows an example for the stereo vision
measurements that were used in [23].
Sensors 2021, 21, 3380 6 of 16

(a) Pixel-wise semantic labeling image (b) Stereo disparity image (c) Stereo disparity confidence image
Figure 2. The three input images to our stereo vision-based grid mapping pipeline used in [23]. © 2021 IEEE. Reprinted,
with permission, from Richter, S.; Beck, J.; Wirges, S.; Stiller, C. Semantic Evidential Grid Mapping based on Stereo Vision.
In Proceedings of the 2020 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems
(MFI), Karlsruhe, Germany, 14–16 September 2020, pp. 179–184, doi: 10.1109/MFI49285.2020.9235217.

2.4. Label Histogram Calculation

As introduced in [1], we calculate the BBA based on the label histogram, which
resembles the contribution of accumulated pixels to a class in a given grid cell. We use a
u-disparity grid and a u-depth grid to compute the label histogram for stereo vision-based
and monocular vision-based grid mapping, respectively. These discretization spaces have
the advantage that disparity and depth estimation uncertainty can be modeled explicitly,
as, e.g., outlined in [8]. For the sake of simplicity, we subsequently refer to the measurement
grid as GM ∈ {Gud , Guz }. The label histogram

hM : G M × S → R≥0

of the measurement M on the u-disparity grid or the u-depth grid is given by

hM (C, ω ) = ∑ wω (C, P) 1{ω } (fsem ( P)),

P∈Guv

where wω is a window function specifying the contribution of the measurement based on

the pixel P ∈ Guv to the cell C ∈ GM and
(
1, if x ∈ X,
1X ( x ) =
0, else,

denotes the indicator function. We apply different measurement models depending on

the assigned semantic hypothesis. For the object hypotheses ω ∈ O , we treat each pixel
measurement P as a point measurement p ∈ P that is the center coordinate of the pixel P.
We then calculate the window function wω based on the inverse sensor model Pr( X ∈ C | p)
to model spatial uncertainty. Here, X denotes the random variable modeling the actual
position that the pixel measurement P is based on. The window function further contains
the confidence fconfk of the corresponding range estimate k ∈ {disp, depth}. Assuming
statistical independence between the spatial uncertainty and the uncertainty of the range
estimate, the window function wω for object classes ω ∈ O is finally set to

wω (C, P) = Pr( X ∈ C | p) · fconfk ( P).

In order to keep the computational complexity at a minimum, we assume X to be

M ( p) ∈ R
uniformly distributed in a rectangle Rc centered around c = Tuv M with size d
such that
µ(C ∩ Rc )
Pr( X ∈ C | p) = ,
µ(Rc )
where µ(·) is the two-dimensional Lebesgue measure. The mapping of pixels with assigned
object labels is sketched in Figure 3.
Sensors 2021, 21, 3380 7 of 16

Guv GM
u

Rc

p∈P
C ∩ Rc

d/z

Figure 3. Mapping of measurements with assigned object labels from image to u-disparity/-
depth grid.

Treating pixels with assigned object labels as points is a simplification based on a lack
of knowledge about object surfaces. For pixels labeled as ground, however, the surface
can be assumed to be locally planar. We use this prior knowledge and propose an ap-
proximating pixel-to-area correspondence to obtain dense mapping results for the ground
hypotheses. The label histogram wω for the ground hypotheses ω ∈ F is given by

1
Z

wω (C, P) = fX ∗ 1AP ( x ) dx, (1)
µ(C )
C

where f X is the probability density function of the random variable X modeling the
measurement position and
ud
A P = Tuv ( P) ⊂ RM
is the area in the grid space corresponding to the measurement pixel P. For pixels classified
with ground labels, the shape of this area depends on the ground surface. We approximate
the resulting label histogram for ground hypotheses by approximating A P with rectangles
based on an encapsulated ground surface estimation in the three steps: ground estimation,
pixel area approximation and area integral calculation.

2.4.1. Ground Surface Estimation

A ground surface estimation is obtained based on the current image measurements in
two stages. First, the height is averaged over all pixels that correspond to a given grid cell
C with an assigned ground label and disparity or depth value exceeding a given threshold.
Here, each pixel is treated as a point measurement leading to sparse mapping results,
especially at far distances. Furthermore, the disparity or depth estimate might add further
sparsity depending on the used method. The quality of the stereo disparity estimation
based on pixel matching, for example, heavily relies on the local contrast of the camera
image. This leads to poor disparity estimation results in smooth areas, especially on the
ground, which results in no height being computed here. The sparse ground estimation
is augmented in the second stage using the inpainting method introduced in [24]. This
inpainting algorithm is based on the Navier–Stokes equations for fluid dynamics and
matches gradients at inpainting region boundaries. To avoid large errors, the inpainting is
only done in a neighborhood of the sparse ground estimation defined by the inpainting
mask fmask . This mask is computed as

fmask ( P) = 1[0,T1 ] (D1 ( P)) · 1[0,T2 ] (D2 ( P)),

Sensors 2021, 21, 3380 8 of 16

where D1 and D2 are the distance transforms based on the masks defining the observed
and unobserved image regions, respectively, and T1 , T2 ∈ R are thresholds defining the
interpolation neighborhood. We justify the application of this data augmentation by the
assumption that gradient jumps in the height profile would lead to gradient jumps in
the pixel intensity and thus imply well-estimated disparity and depth. Consequently,
the inpainted regions are restricted to areas without jumps in the height profile.

2.4.2. Pixel Area Approximation

Depending on the ground surface relief, a pixel patch may correspond to arbitrarily
shaped areas A P in the grid space. Given a grid cell C ∈ Gud , the label histogram from
Equation (1) for the ground label ω ∈ F contains the sum of all pixel portions overlapping
with the cell projected into the image domain. To accelerate the mapping process, we
approximate the projected cell by a rectangle R P . Utilizing the estimated ground height,
the homogeneous position in cartesian coordinates c0 = ( x, y, z, 1)T can be computed based
on the lower left cell corner point cll and the upper right cell corner point cur in u-disparity
space. The lower left sub-pixel (ull , vll ) and the upper right sub-pixel coordinates (uur , vur )
are then computed using the perspective mapping F as

(ui , vi , di ) = F (ci0 ) = K · ci0 , i ∈ {ll, ur },

where K is the 3 × 4 pinhole camera matrix. Based on the projected corner points, the pro-
jected rectangle is then given by

R P = [ull , uur ) × [vll , vur ).

The grid cell approximation is depicted in Figure 4.

Guv GM
u
v

AP
C

RP

d/z

Figure 4. Mapping of measurements with assigned ground labels from image to u-disparity/-
depth grid.

2.4.3. Area Integral Calculation

Finally, the label histogram approximation for the ground labels ω ∈ F can efficiently
be calculated based on the integral image

k l
f̄ω ( I1,k , I2,l ) = ∑ ∑ 1{ω} (fsem ( I1,i , I2,j ))
i =0 j =0

as

hM (c, ω ) ≈ f̄ω (uur , vur ) − f̄ω (uur , vll )

− f̄ω (ull , vur ) + f̄ω (ull , vll ).
Sensors 2021, 21, 3380 9 of 16

Note that in the upper equation, ui and vi are sub-pixel coordinates and the corre-
sponding integral image f̄ω is evaluated using bilinear interpolation.

2.5. Grid Transformation to Cartesian Space

The label histogram grid map layers hM (·, ω ) are transformed from u-disparity/-
depth space to the cartesian space before the BBA calculation to prevent inconsistencies in
the belief assignment due to interpolation artifacts. Note that cartesian grid cells close to
the camera, for instance, correspond to many u-disparity grid cells, while one u-disparity
grid cell covers several cartesian grid cells at far distances. The relations between the
considered tessellations are sketched in Figure 5. Non-regular cell area correspondences
occur not only between u-disparity/-depth and cartesian grids. Yguel et al. investigated
this effect in detail for the switch from a polar to cartesian coordinate system in [25]. Simple
remapping methods lead to the so-called Moiré effect due to undersampling, which is
well known in computer graphics. We tackle this issue by applying a well-established
upsampling principle in relevant areas. By analyzing the area ratio between a cartesian
cell and the corresponding u-disparity/-depth cell, a set of points is chosen lying on an
equidistant grid within the cell. The u-disparity/-depth coordinate is calculated for each
point, and the label histogram value is computed based on the u-disparity/-depth grid
map utilizing bilinear interpolation. The label histogram’s final value for the cartesian cell
is a weighted average over all sampled cell points.
xy
Tud
Gud G xy

d x

u y
xy
Tuz
Guz G xy

z x

u y

Figure 5. In the cartesian grid on the right-hand side, the grid cells are influenced by the distorted
overlayed areas based on the corresponding u-disparity or u-depth grid cell, respectively.
Sensors 2021, 21, 3380 10 of 16

2.6. Basic Belief Assignment

The label histogram is subsequently used to compute a consistent BBA in the mea-
surement grid map gM . The BBA is computed based on the false-positive probability
pω as  
∏
h (C,θ ) h (C,ω )
gM (C, ω ) = pθ M 1 − pωM ,
θ ∈Ω\{ω }

for the relevant hypotheses ω ∈ S . Note that pω can easily be determined based on the
confusion matrix of the evaluation data set of the semantic labelling network.

3. Results
We execute our proposed method based on two setups using the Kitti odometry
benchmark [26]. In the first case, we calculate stereo disparities based on the two color
cameras in the Kitti sensor setup using the guided aggregation net for stereo matching
presented by Zhang et al. in [27]. The authors connect a local guided aggregation layer
that follows a traditional cost filtering refining thin structures to a semi-global aggregation
layer. In the second setup, we only use the left color camera and compute a depth map
using the unsupervised method presented by Godard et al. in [21]. Both neural networks
are openly available on GitHub and have been trained or at least refined using the Kitti
2015 stereo vision benchmark. For calculating the pixelwise semantic labeling, the neural
network proposed by Zhu et al. in [28] was used. Their network architecture is openly
available as well and achieves a mean intersection over union (IoU) of 72.8% in the Kitti
semantic segmentation benchmark. Note that all of the above choices were made inde-
pendently of runtime considerations. In both cases, the pixelwise confidences for depth
and disparity, respectively, are set to one as the corresponding networks do not output
adequate information. In Figure 6, an example of the three used input images is depicted.
Semantic segmentation fsem

Stereo disparities fdisp Depth from mono fdepth

Figure 6. Results of the three neural networks used to generate the input images for our proposed
grid mapping pipeline.

The region of interest of our cartesian grid map is 100 m in x-direction and 50 m
in y-direction where the sensor origin is located at (0 m, 25 m). The cell size is 10 cm in
both dimensions.
In the remainder of this section, we first present the ground truth that we used to
evaluate our method in Section 3.1. We then present some visual results in Section 3.2.
Finally, we present a detailed quantitative evaluation in Section 3.3.

3.1. Ground Truth Generation

We base our quantitative evaluation on the SemanticKITTI dataset presented by
Behley et al. in [29]. SemanticKitti extends the Kitti odometry benchmark by annotating the
360◦ Lidar scans with semantics labels using a set of 28 classes. Here, we merge those classes
to obtain semantic labels corresponding to our singleton hypotheses S . Using the labeled
poses in the Kitti odometry dataset, the point clouds from ten frames are transformed
into the current pose, compensating for ego-motion and the subsequent accumulation.
This densifies the semantic point cloud around the ego vehicle. The thus accumulated
Sensors 2021, 21, 3380 11 of 16

3D semantically annotated point cloud is mapped into the same grid G xy that is used for
our semantic evidential grid map representation. The generation of our ground truth is
illustrated in Figure 7. When using multiple frames to build a denser ground truth, some
cells covered by dynamic objects are covered by road pixels. In order to remove those
conflicts, we use morphological operations to remove the ground labels in those regions, as
can be seen in Figure 7 at the locations of the two vehicles that are present in the depicted
scene. Subsequently, the grid map containing the ground truth labels is denoted as

gGT : G xy → S ∪ {unknown},

assigning both a semantic label or the label “unknown” to each grid cell C ∈ G xy .

SemanticKitti 3D point cloud 10 frames

topview

Figure 7. The generation of the ground truth used for the quantitative evaluation. Three-dimensional
semantic point clouds from ten frames are merged and mapped into a top-view grid.

3.2. Visual Evaluation

We process the first 1000 frames for the sequences 00 to 10 in the Kitti odometry
dataset. Figure 8 depicts visual impressions of the results for the sequences 00, 01, and 02.
The first thing that stands out is that the detection range in both the mono- and stereo-based
grid maps surpasses the one in the Lidar-based ground truth. The BBA in our resulting
evidential grid maps decreases with the distance to the sensor origin, which aligns with
the intuition that the uncertainty is higher at larger distances. The first scenario in the left
column was captured in a suburban region in Karlsruhe and contains a series of residences
on the left, a t-crossing to the left, and a sidewalk on the right that is separated from the
road by terrain. The border of the residences appears to be captured better when using
the stereo pipeline. As the sidewalk on the right is covered by shadows leading to low
contrast in the corresponding image region, its geometry cannot be captured with both
pipelines. The middle column shows a highway scenario with a vehicle in front of the ego
vehicle at about 50 m distance. The guardrails in both the ego vehicle and the opposite
lane can only be captured using the stereo pipeline. The leading vehicle is detected more
precisely using the stereo pipeline as well. In the third column, a scenario in a rural area
with a vehicle passing on the opposite lane is depicted. There is a sidewalk on each side of
the road with adjacent terrain. The rough geometry of the parts can be captured in both
Sensors 2021, 21, 3380 12 of 16

the mono and the stereo pipeline. The passing vehicle is detected better using the stereo
pipeline, whereas its shape is slightly distorted using the monocular vision pipeline due to
higher inaccuracy in the depth estimation. As a general observation, it can be stated that
the errors in both camera-based reconstructions are dominated by flying pixels at object
boundaries that result from inconsistencies between the pixelwise semantic estimate and
the depth or disparity estimation.

Figure 8. The resulting BBA for stereo and mono images. Each column corresponds to one frame in the Kitti odometry
benchmark depicted in the image in the first row. The second row shows the ground truth, the third row shows the results
for stereo vision, and the last row shows the results for monovision.

3.3. Quantitative Evaluation

We provide a quantitative evaluation of our method based on the intersection over
union per class and the overall ratio of correctly predicted cell states.

3.3.1. Intersection over Union

The intersection over union (IoU) and the mean intersection over union (mIoU) are
defined as
TPω 1
|S| ω∑
IoUω = , mIoU = IoUω ,
TPω + FPω + FNω ∈S

where TPω presents the number of true positive cells, FPω the number of false-positive
cells, and FNω the number of false-negative cells of the label ω ∈ S . In this context, a
grid cell is considered as a true positive if the class in the ground truth coincides with the
class ω ∈ S that has been assigned the highest BBA. Note, hence, that this metric does not
Sensors 2021, 21, 3380 13 of 16

consider the measure of uncertainty that is encoded in the BBA. Therefore, we calculate the
modified intersection-over-union metrics

TP0ω 1
|S| ω∑
IoU0ω = , mIoU0 = IoU0ω ,
TP0ω + FP0ω + FN0ω ∈S

based on the modified rates

TP0ω = ∑ 1{ gGT (C)} (ω ) gM (C, ω ),

C ∈G xy

FP0ω = ∑ 1{S\ gGT (C)} (ω ) gM (C, ω ), FN0ω = ∑ FP0φ .

C ∈G xy φ∈S\ω

Tables 1 and 2 show the above-defined IoU metrics for the sequences 00 to 10 in the
Kitti odometry benchmark. The tables contain the numbers for all considered semantic
classes except for other movable objects (om) as it barely occurs in the test sequences. The
stereo vision pipeline outperforms the monocular vision pipeline for almost all classes.
This is expected as the used stereo disparity estimation is more accurate than the monocular
depth estimation. In general, the numbers for both setups are in similar regions as the ones
presented in the Lidar-based semantic grid map estimation from Bieder et al. in [30]. They
reach a 39.8% mean IoU with their best configuration. Our proposed method reaches 37.4%
and 41.0% mean IoU in the monocular and stereo pipeline, respectively. The accuracy for
small objects as pedestrians and cyclists is very low as small errors in the range estimations
have a high effect compared to the objects size. Comparing the numbers for mIoUω with
mIoU0ω incorporating the BBA, it stands out that the modified IoU is significantly higher.
For the modified IoU, means of 44.7% and 48.7% are reached in the two setups. The
reason for this is that higher uncertainties in wrongly classified cells lower the modified
false-positive and false-negative rates FP0ω and FN0ω and thus also the denominator in
the calculation of mIoU0ω . The results show that wrong classifications are attached with a
higher uncertainty and that the BBA can be used as a meaningful measure for uncertainty.

Table 1. Class IoUs IoUω (IoU0ω ) for the stereo vision pipeline in %. The dash indicates that there are no corresponding
objects in the sequence. The column on the right contains the mean IoUs mIoU (mIoU0 ).

Seq. Car Cyclist Pedestrian Non Movable Street Sidewalk Terrain ∅

00 51.0 (65.7) 5.4 (6.3) 3.4 (4.7) 40.6 (50.9) 92.3 (95.6) 64.4 (72.7) 29.0 (35.3) 40.9 (47.3)
01 22.8 (45.8) 10. 3(8.8) - 27.0 (36.6) 85.3 (92.8) - 59.2 (66.5) 29.2 (35.8)
02 48.9 (66.7) 3.0 (2.9) 0.3 (0.3) 17.4 (23.8) 86.5 (91.4) 49.8 (60.3) 56.2 (57.0) 37.4 (43.2)
03 33.8 (54.1) 2.0 (3.2) - 26.2 (33.6) 84.7 (88.4) 60.5 (67.0) 82.6 (85.2) 41.4 (47.4)
04 45.7 (64.0) - - 26.7 (31.0) 90.1 (92.9) 34.9 (43.7) 64.2 (67.4) 37.4 (42.7)
05 43.6 (60.4) 3.3 (5.2) 6.0 (8.2) 32.4 (43.3) 88.8 (93.1) 57.7 (67.5) 20.0 (23.5) 36.0 (43.0)
06 31.7 (49.1) 4.2 (4.9) 1.5 (1.6) 28.0 (39.8) 80.8 (88.9) 50.2 (62.3) 79.1 (84.4) 39.4 (47.3)
07 44.2 (61.6) 5.7 (7.1) 15.6 (17.8) 43.0 (52.3) 89.3 (93.7) 61.9 (69.4) 71.7 (76.2) 47.3 (54.0)
08 37.5 (56.1) 8.3 (12.4) 6.6 (10.2) 33.5 (46.7) 87.1 (91.9) 57.0 (67.1) 72.1 (75.2) 43.2 (51.4)
09 37.7 (57.7) 5.0 (6.1) 5.7 (15.7) 29.3 (41.7) 85.4 (90.7) 53.4 (64.9) 60.3 (65.7) 39.5 (49.0)
10 33.4 (50.8) - 4.6 (7.2) 28.2 (36.4) 80.6 (85.2) 45.3 (52.5) 48.8 (53.5) 34.4 (40.8)
all 40.8 (59.0) 5.0 (6.7) 5.5 (8.1) 30.7 (41.4) 85.9 (91.3) 54.2 (64.2) 65.1 (69.9) 41.0 (48.7)
Sensors 2021, 21, 3380 14 of 16

Table 2. Class IoUs IoUω (IoU0ω ) for the monocular vision pipeline in %. The dash indicates that there are no corresponding
objects in the sequence. The column on the right contains the mean IoUs mIoU (mIoU0 ).

Seq. Car Cyclist Pedestrian Non Movable Street Sidewalk Terrain ∅

00 40.6 (51.5) 3.5 (3.6) 0.4 (0.7) 42.4 (59.1) 89.9 (92.8) 60.4 (71.2) 30.7 (43.3) 38.3 (46.0)
01 15.4 (25.0) 7.1 (9.0) - 18.3 (13.1) 82.8 (90.9) - 63.4 (72.6) 26.7 (30.1)
02 35.6 (50.4) 3.2 (3.9) 0.1 (0.1) 16.1 (22.6) 84.7 (88.9) 46.8 (54.9) 56.9 (60.8) 34.8 (40.2)
03 21.9 (31.7) 2.1 (3.4) - 16.9 (18.9) 79.6 (83.8) 49.2 (55.9) 75.6 (81.8) 35.0 (39.4)
04 15.6 (30.6) - - 31.6 (36.4) 86.9 (90.3) 35.0 (41.2) 65.1 (70.6) 33.5 (38.4)
05 25.5 (39.5) 4.5 (5.6) 1.6 (2.6) 31.4 (45.6) 86.9 (90.4) 54.2 (65.1) 20.3 (34.1) 32.1 (40.4)
06 19.2 (30.4) 4.0 (5.0) 1.0 (1.7) 32.3 (49.0) 77.8 (84.8) 46.7 (57.9) 79.6 (86.0) 37.2 (45.0)
07 32.5 (46.5) 5.3 (6.4) 4.3 (7.4) 44.0 (57.9) 85.9 (90.3) 57.6 (68.2) 70.5 (81.0) 42.9 (51.1)
08 24.9 (37.8) 6.7 (8.8) 1.6 (2.0) 35.8 (54.2) 85.0 (89.1) 53.4 (62.8) 71.0 (77.9) 39.8 (47.5)
09 26.1 (40.8) 2.6 (2.7) 3.8 (7.1) 32.9 (44.9) 83.5 (88.4) 49.5 (60.6) 62.4 (72.1) 37.3 (45.2)
10 19.9 (31.7) - 3.0 (3.9) 25.2 (34.4) 76.2 (80.6) 40.7 (48.1) 47.4 (61.2) 30.4 (37.1)
all 27.7 (41.2) 4.2 (5.2) 2.0 (3.0) 29.5 ( 41.7) 83.2 (88.3) 49.9 (59.6) 65.0 (74.1) 37.4 (44.7)

3.3.2. Ratio of Correct Labels

As a second class of metrics, we consider the ratio of correctly classified cells

∑C∈Gxy TC
CR = ,
∑C∈Gxy (TC + FC )

where TC ∈ {0, 1} equals one if the correct label ω ∈ S was assigned the highest BBA
greater than zero and FC ∈ {0, 1} is one if the highest BBA greater than zero corresponds
to the wrong label. The counterpart incorporating the BBA reads

∑C∈Gxy TC mC
CR0 = , mC = arg max gM (C, ω ).
∑C∈Gxy (TC + FC )mC ω ∈S

We have calculated CR0 as well as CR for sequences 00 to 10. The results are presented
in Table 3 for the stereo vision pipeline and in Table 4 for the monocular vision pipeline.
The numbers confirm the tendencies collected in the IoU-based evaluation. The modified
ratios CR0 based on the BBA are higher than the ones that are based solely on one predicted
class per cell and the ratios of the stereo vision pipeline are slightly above the ones of the
monocular vision pipeline. Besides the consistency between range estimation and semantic
segmentation, the quality of the semantic segmentation itself naturally influences the final
results strongly. We found that the majority of the errors in the segmentation occur in
the distinction between the road and the sidewalk. Experiments showed that CR0 can be
improved by up to 10% depending on the sequence when merging the two classes. Besides
the Lidar-based semantic top-view maps presented in [30], we can compare our results to
the hybrid approach using Lidar and RGB images from Erkent et al. presented in [14]. They
achieve a ratio of correctly labeled cells of 81% in their best performing setup, indicating
that our approach performs slightly better. However, note that they predict a different set
of classes without uncertainty considerations.

Table 3. Ratio of correctly labeled grid cells for the stereo vision pipeline.

Seq. 00 01 02 03 04 05 06 07 08 09 10 All
CR 81.5 81.1 78.6 86.7 83.7 73.6 82.5 83.9 82.9 79.3 73.4 80.8
CR0 87.0 87.9 84.2 89.3 87.9 81.3 88.2 88.4 87.8 85.9 78.8 86.2
Sensors 2021, 21, 3380 15 of 16

Table 4. Ratio of correctly labeled grid cells for the monocular vision pipeline.

Seq. 00 01 02 03 04 05 06 07 08 09 10 All
CR 80.3 81.0 77.7 81.5 83.1 71.9 81.9 81.6 81.5 78.9 70.1 79.3
CR0 87.8 88.9 83.8 86.3 87.7 83.4 88.1 88.3 87.5 86.2 78.5 86.3

4. Conclusions
We presented an accurate and efficient framework for semantic evidential grid map-
ping based on two camera setups: monocular vision and stereo vision. Our resulting
top-view representation contains evidential measures for eight semantic hypotheses, which
can be seen as a refinement of the classical occupancy hypotheses free and occupied. We
explicitly model uncertainties of the sensor setup-dependant range estimation in an inter-
mediate grid representation. The mapping results are dense and smooth, yet not complete
as no estimates are given in unobserved areas. In our quantitative evaluation, we showed
the benefits of our evidential model by obtaining significantly better error metrics when
considering the uncertainties. This is one of the main advantages of our method compared
to other publications and enables our pipeline to perform comparably well to competitive
ones using more expensive sensors such as Lidar [14,30]. The second advantage is the
underlying semantic evidential representation that makes fusion with other sensor types
as range sensors straight forward, see [1]. The main bottlenecks in our pipeline are the
semantic segmentation and the range estimation in the image domain as well as the consis-
tency between both. Especially the influence of the latter might easily be underestimated
as inconsistencies of a few pixels already imply large distortions at higher distances.
In future work, we will focus on developing a refinement method to improve the
consistency between range and semantic estimation in the image domain. In this regard,
it might also be promising to combine both in a mutual aid network to achieve a higher
consistency in the first place. We will then fuse the presented vision-based semantic
evidential grid maps with evidential grid maps from range sensors based on the method
described in [1]. Furthermore, we will incorporate the fused grid maps into a dynamic
grid mapping framework that is able to both accumulate a semantic evidential map as
well as track dynamic traffic participants. Finally, we aim at providing a real-time capable
implementation of our framework by utilizing massive parallelization on state-of-the-
art GPUs.

Author Contributions: Conceptualization, S.R., J.B. and S.W.; methodology, S.R., J.B.; software,
S.R., Y.W.; validation, S.R., J.B., S.W. and Y.W.; formal analysis, S.R. and Y.W.; investigation, S.R.
and Y.W.; resources, S.R.; data curation, S.R. and Y.W.; writing—original draft preparation, S.R.
and Y.W.; writing—review and editing, S.R.; visualization, S.R. and Y.W.; supervision, C.S.; project
administration, C.S.; funding acquisition, S.R. All authors have read and agreed to the published
version of the manuscript.
Funding: We acknowledge the support from the KIT-Publication Fund of the Karlsruhe Institute
of Technology.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Richter, S.; Wirges, S.; Königshof, H.; Stiller, C. Fusion of range measurements and semantic estimates in an evidential
framework/Fusion von Distanzmessungen und semantischen Größen im Rahmen der Evidenztheorie. Tm-Tech. Mess. 2019,
86, 102–106. [CrossRef]
2. Shafer, G. A Mathematical Theory of Evidence; Princeton University Press: Princeton, NJ, USA, 1976; Volume 42.
3. Elfes, A. Using Occupancy Grids for Mobile Robot Perception and Navigation. Computer 1989, 22, 46–57. [CrossRef]
4. Nuss, D.; Reuter, S.; Thom, M.; Yuan, T.; Krehl, G.; Maile, M.; Gern, A.; Dietmayer, K. A Random Finite Set Approach for Dynamic
Occupancy Grid Maps with Real-time Application. Int. J. Robot. Res. 2018, 37, 841–866. [CrossRef]
5. Steyer, S.; Tanzmeister, G.; Wollherr, D. Grid-Based Environment Estimation Using Evidential Mapping and Particle Tracking.
IEEE Trans. Intell. Veh. 2018, 3, 384–396. [CrossRef]
Sensors 2021, 21, 3380 16 of 16

6. Badino, H.; Franke, U. Free Space Computation Using Stochastic Occupancy Grids and Dynamic Programming; Technical Report;
Citeseer: University Park, PA, USA, 2007.
7. Perrollaz, M.; Spalanzani, A.; Aubert, D. Probabilistic Representation of the Uncertainty of Stereo-Vision and Application to
Obstacle Detection. In Proceedings of the IEEE Intelligent Vehicles Symposium, La Jolla, CA, USA, 21–24 June 2010; pp. 313–318.
[CrossRef]
8. Pocol, C.; Nedevschi, S.; Meinecke, M.M. Obstacle Detection Based on Dense Stereovision for Urban ACC Systems. In Proceedings
of the 5th International Workshop on Intelligent Transportation, Hamburg, Germany, 18–19 March 2008.
9. Danescu, R.; Pantilie, C.; Oniga, F.; Nedevschi, S. Particle Grid Tracking System Stereovision Based Obstacle Perception in Driving
Environments. IEEE Intell. Transp. Syst. Mag. 2012, 4, 6–20. [CrossRef]
10. Yu, C.; Cherfaoui, V.; Bonnifait, P. Evidential Occupancy Grid Mapping with Stereo-Vision. In Proceedings of the IEEE Intelligent
Vehicles Symposium, Seoul, Korea, 28 June–1 July 2015; pp. 712–717. [CrossRef]
11. Giovani, B.V.; Victorino, A.C.; Ferreira, J.V. Stereo Vision for Dynamic Urban Environment Perception Using Semantic Context
in Evidential Grid. In Proceedings of the IEEE Conference on Intelligent Transportation Systems, Gran Canaria, Spain, 15–18
September 2015; pp. 2471–2476. [CrossRef]
12. Valente, M.; Joly, C.; de la Fortelle, A. Fusing Laser Scanner and Stereo Camera in Evidential Grid Maps. In Proceedings of the
2018 15th International Conference on Control, Automation, Robotics and Vision (ICARCV), Singapore, 18–21 November 2018.
13. Thomas, J.; Tatsch, J.; Van Ekeren, W.; Rojas, R.; Knoll, A. Semantic grid-based road model estimation for autonomous driving. In
Proceedings of the 2019 IEEE Intelligent Vehicles Symposium (IV), Paris, France, 9–12 June 2019. [CrossRef]
14. Erkent, O.; Wolf, C.; Laugier, C.; Gonzalez, D.S.; Cano, V.R. Semantic Grid Estimation with a Hybrid Bayesian and Deep Neural
Network Approach. In Proceedings of the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS),
Madrid, Spain, 1–5 October 2018; pp. 888–895.
15. Lu, C.; van de Molengraft, M.J.G.; Dubbelman, G. Monocular Semantic Occupancy Grid Mapping With Convolutional Variational
Encoder–Decoder Networks. IEEE Robot. Autom. Lett. 2019, 4, 445–452. [CrossRef]
16. Eigen, D.; Puhrsch, C.; Fergus, R. Depth map prediction from a single image using a multi-scale deep network. Advances in
neural information processing systems. arXiv 2014, arXiv:1406.2283.
17. Liu, F.; Shen, C.; Lin, G.; Reid, I. Learning depth from single monocular images using deep convolutional neural fields. IEEE
Trans. Pattern Anal. Mach. Intell. 2015, 38, 2024–2039. [CrossRef] [PubMed]
18. Fu, H.; Gong, M.; Wang, C.; Batmanghelich, K.; Tao, D. Deep ordinal regression network for monocular depth estimation. In
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018;
pp. 2002–2011.
19. Alhashim, I.; Wonka, P. High quality monocular depth estimation via transfer learning. arXiv 2018, arXiv:1812.11941.
20. Zhou, T.; Brown, M.; Snavely, N.; Lowe, D.G. Unsupervised learning of depth and ego-motion from video. In Proceedings of the
IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 July 2017; pp. 1851–1858.
21. Godard, C.; Mac Aodha, O.; Brostow, G.J. Unsupervised monocular depth estimation with left-right consistency. In Proceedings
of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 July 2017; pp. 270–279.
22. Qiao, S.; Zhu, Y.; Adam, H.; Yuille, A.; Chen, L.C. ViP-DeepLab: Learning Visual Perception with Depth-aware Video Panoptic
Segmentation. arXiv 2020, arXiv:2012.05258.
23. Richter, S.; Beck, J.; Wirges, S.; Stiller, C. Semantic Evidential Grid Mapping based on Stereo Vision. In Proceedings of the 2020
IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI), Karlsruhe, Germany, 14–16
September 2020; pp. 179–184. [CrossRef]
24. Bertalmio, M.; Bertozzi, A.; Sapiro, G. Navier-Stokes, Fluid Dynamics, and Image and Video Inpainting. In Proceedings of
the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2001, Kauai, HI, USA, 8–14
December 2001; Volume 1, pp. 1–355. [CrossRef]
25. Yguel, M.; Aycard, O.; Laugier, C. Efficient GPU-based construction of occupancy grids using several laser range-finders. Int. J.
Veh. Auton. Syst. 2008, 6, 48–83. [CrossRef]
26. Geiger, A.; Lenz, P.; Stiller, C.; Urtasun, R. Vision meets robotics: The kitti dataset. Int. J. Robot. Res. 2013, 32, 1231–1237.
[CrossRef]
27. Zhang, F.; Prisacariu, V.; Yang, R.; Torr, P.H.S. GA-Net: Guided Aggregation Net for End-To-End Stereo Matching. In Proceedings
of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 16–20 June 2019;
pp. 185–194. [CrossRef]
28. Zhu, Y.; Sapra, K.; Reda, F.A.; Shih, K.J.; Newsam, S.; Tao, A.; Catanzaro, B. Improving semantic segmentation via video
propagation and label relaxation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long
Beach, CA, USA, 16–20 June 2019; pp. 8856–8865.
29. Behley, J.; Garbade, M.; Milioto, A.; Quenzel, J.; Behnke, S.; Stachniss, C.; Gall, J. SemanticKITTI: A Dataset for Semantic Scene
Understanding of LiDAR Sequences. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV),
Seoul, Korea, 27 October–2 November 2019.
30. Bieder, F.; Wirges, S.; Janosovits, J.; Richter, S.; Wang, Z.; Stiller, C. Exploiting Multi-Layer Grid Maps for Surround-View Semantic
Segmentation of Sparse LiDAR Data. arXiv 2020, arXiv:2005.06667.