• Aucun résultat trouvé

Scene understanding : Toward a Safer Navigation

N/A
N/A
Protected

Academic year: 2021

Partager "Scene understanding : Toward a Safer Navigation"

Copied!
52
0
0

Texte intégral

(1)

an author's https://oatao.univ-toulouse.fr/21293

Vivet, Damien Scene understanding : Toward a Safer Navigation. (2018) In: ICUTS, 9 July 2018 - 12 July 2018 (Edmonton, Canada). (Unpublished)

(2)

SCENE

UNDERSTANDING:

Toward a Safer Navigation

Damien VIVET ISAE-SUPAERO • DEOS

(3)

A three years old children is an expert of image analysis, content description and event recognition.

(4)

Our society is more technologically advanced than ever but still less

advanced than a 3 years old child.

(5)
(6)
(7)
(8)

We already have excellent sensors

(9)
(10)

One mission:

To give computers

visuals intelligence.

(11)

1800 lecturers from the industry

200 professors and researchers

43 research directors (HDR)

21500 Alumnis / 1700 students

30% of international students

55 nationalities

"Ingénieur ISAE-SUPAERO" (MsC)

15 Advanced Masters

6 Doctoral programs

(12)

To navigate we need three major capabilities :

- perception - command law

(13)

The visual perception: Man VS Machine

(14)
(15)

Humans also use landmarks to localize

(16)

First localization and mapping based on landmarks

(17)

Modern

(18)

A landmark for a computer:

(19)

Applications of multi-view geometry

“Building Rome in a Day”, S. Agarwal, N. Snavely, I. Simon, S. M. Seitz and R. Szeliski

(20)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(21)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(22)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(23)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(24)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(25)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(26)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(27)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(28)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(29)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(30)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(31)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(32)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(33)

“Simultaneous Localization

and Mapping: SLAM”

SLAM is a process in which a vehicle build a representation of its environment while localizing

(34)

“Simultaneous Localization

and Mapping: SLAM”

ISAE-SUPAERO, France:

(35)

“Simultaneous Localization and Mapping:

SLAM”

(36)

“Mobile object detection

and tracking: DATMO”

(37)

“Traffic monitoring”

LITIS (INSA Rouen), France and ISAE-SUPAERO Dr. Yadu Prabhakar (NAIT Edmonton)

(38)

"What to do next?”

How to go from standard “geometry” to “semantics”? We just have to be smarter...

(39)
(40)

"Road infrastructure detection

and mapping”

(41)

"Road infrastructure detection and

mapping”

(42)

"Road infrastructure detection and

mapping”

(43)

"Road infrastructure detection and

mapping”

(44)

"Road infrastructure detection and

mapping”

(45)

"Road infrastructure detection and

mapping”

(46)

"Localization based on Road

infrastructure detection”

(47)

"Human behavior detection”

(48)

"Abnormality detection”

(49)
(50)

"How to go deeper ?”

A neural network is a simple assembly of connected elements called “perceptron”.

The most you have, the Deeper you are.

(51)

"Road scene labelling”

Chinese University of Hong Kong University Paris-Est

(52)

“Merci pour votre attention !!!”

Des questions ?

Damien Vivet

Références

Documents relatifs