Public
Point Cloud Seq
Segmentation
42%
The first 12 SemanticKITTI sequences
  • Dec 9, 2021
  • 12 samples
ReadMe

SemanticKITTI is a dataset of LiDAR sequences for semantic scene understanding - more info on http://www.semantic-kitti.org/

The labels in this dataset come from the panoptic segmentation task: http://www.semantic-kitti.org/tasks.html#panseg

Labeling specification
id name description attributes
1outlier - yes
10car - yes
11bicycle - -
13bus - -
15motorcycle - -
16on-rails - -
18truck - -
20other-vehicle - -
30person - -
31bicyclist - -
32motorcyclist - -
40road - -
44parking - -
48sidewalk - -
49other-ground - -
50building - -
51fence - -
52other-structure - -
60lane-marking - -
70vegetation - -
71trunk - -
72terrain - -
80pole - -
81traffic-sign - -
99other-object - -
252moving-car - -
253moving-bicyclist - -
254moving-person - -
255moving-motorcyclist - -
256moving-on-rails - -
257moving-bus - -
258moving-truck - -
259moving-other-vehicle - -
Image-level attributes
None
Labeling instructions
Label statistics
Last updated 2 years ago ago.
car
2253
moving-car
487
outlier
313
road
300
fence
300
vegetation
300
pole
300
terrain
294
traffic-sign
264
sidewalk
250
trunk
250
other-object
250
other-vehicle
203
parking
200
building
200
moving-person
179
other-ground
132
other-structure
131
truck
107
moving-motorcyclist
100
bus
55
person
55
lane-marking
50
moving-bicyclist
23
motorcyclist
2
bicycle
1
motorcycle
0
on-rails
0
bicyclist
0
moving-on-rails
0
moving-bus
0
moving-truck
0
moving-other-vehicle
0