benjamin schiereck, sales manger traficon germany
Post on 17-Jan-2016
37 Views
Preview:
DESCRIPTION
TRANSCRIPT
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Benjamin Schiereck, Sales Manger Traficon Germany
Improving road and tunnel Improving road and tunnel safety via incident safety via incident management:management: implementing a video image implementing a video image processing systemprocessing system
Vid
eo D
etec
tion
Sol
utio
nsV
ideo
Det
ectio
n S
olut
ions
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
• Introduction• Incident Management: Video Based Incident
Detection Basic Functions of Incident Management Video Image Processing Functions, Methodology & System
architecture Detection rate Automatic Incident Detection system Cases (Eye on fire in a tunnel) Typical Freeway and Tunnel AVI files
• Summary• Conclusions
Outline of PresentationOutline of Presentation
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
More Traffic
More Accidents
&
More Cars
Involve
d
Time
More Secondary
Accidents
&
Long Traffic Ja
ms
SOLUTION?
INCID
ENT
INCID
ENT
MANAGEMENT
MANAGEMENT
IntroductionIntroduction
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
1. Traffic Monitoring, Prevention
2. Incident Detection
3. Incident Verification
4. Driver Information
5. Incident Clearing
Basic Functions of Incident ManagementBasic Functions of Incident Management
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Traffic Monitoring – Prevention Traffic Monitoring – Prevention
• Most important is safe infrastructure
• Monitor traffic situation, speed and occupancy using video cameras.
• Set appropriate speeds on VMS panels
• Fast information about the incident
• Fast reaktion on incident– e.G.. Closing the Tunnel
All about time
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Traffic Monitoring on HighwayTraffic Monitoring on Highway
using cameras!!
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Traffic Monitoring in TunnelTraffic Monitoring in Tunnel
• Access control, situation in the Tunnel• Monitoring actions with video
1. Slow driving vehicle
2. Traffic jam in tunnel
3. Speed differences
4. Occupancy
5. Intervehicle distances
• VMS Panels• Ventilation control
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Detection RateDetection RateIncident Detection with respect to dedicated camera positions for incident detection
Indoors (tunnel)
Outdoors Time to detect
stopped vehicles (%) 98 95 10 sec
queue (%) 99,9 99,5 2 sec
inverse direction (%) 95 95 < 1 sec
flow speed (maximal error) 10 % 10 %
false alarm frequency(per camera / per day)
0,025 0,15
Data collection for Outdoors Applications and for dedicated camera positions for data collection
Counting
Speed Queue
> 98 % > 95% with errors < 5% 99%
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
1. Importance to avoid traffic jams
2. Importance to avoid secondary accidentsVerona, ITALY Foix, FRANCE
Direct Incident detection,Time to Detect, Time to VerifyDirect Incident detection,Time to Detect, Time to Verify
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Verona, ITALY
Incident ManagementVideo Based Incident DetectionIncident ManagementVideo Based Incident Detection
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
• Stopped vehicle• Slow moving• Counting• Inverse direction• Distances between cars• Fallen objects• Pedestrians• Smoke
Video Image Processing FunctionsVideo Image Processing Functions
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
System Architecture System Architecture
CAMERA VIP T M S
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Case 1: Fire in a Tunnel – Oslo 1996Case 1: Fire in a Tunnel – Oslo 1996
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Data from Escota France,1999
Type of Vehicle
Stopped Vehicle
Visible Smoke
First VisibleFlames
Global Fire
Car 0 min. 3 min. 5 min. 8 min.
Van 0 min. 5 min. 8 min. 15 min.
Engine fire (2%)
0 min. Fast Fast Fast
Brake Fire (98%)
0 min. 10 min. 12 min. 20 min.
Evolution of Fires of Vehicles in and around Tunnels
Video Image Processing MethodologyVideo Image Processing Methodology
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
• Because this is quite a long tunnel at under sea level, the owner requested a highly redundant system with a very high detection rate, high reliability (MTBF) and a very high level of service (% Uptime).
• This was one of the reasons why the detection cameras were installed at 60 metres distance, but programmed to cover at least 120 m.
Case 2: ORESUND - SituationCase 2: ORESUND - Situation
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
• Detection Rate• False Detections• False Detection Cost• Reliability
Öresund
ORESUND : Other considerationsORESUND : Other considerations
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
100m 100m 100m 100m 100mC1 C2 C3 C4 C5 C6
100m
FigureFigure11: Distance between two cameras set at 100 metres without : Distance between two cameras set at 100 metres without overlapping field of viewoverlapping field of view
FigureFigure 2 2: Distance between two cameras set at 60 metres with : Distance between two cameras set at 60 metres with overlapping field of view overlapping field of view
C1 C2 C3 C4 C5 C660m 60m 60m 60m 60m
120m
60m 60m
RedundancyRedundancy
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Video Detection examples: Tunnel ApplicationsVideo Detection examples: Tunnel Applications
Smoke Detection
Object Detection
Pedestrian Detection
Inverse Direction Detection
Incident Detection
Stopped Vehicle Detection
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Video Detection examples: Highway & Bridge ApplicationsVideo Detection examples: Highway & Bridge Applications
Inverse Direction Detection at night
Stopped Vehicle Detection
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Basic Advantages of Video Based Incident Detection:
• Fast incident detection rate• Visual verification • High system reliability• Easy to install and modify• Low false detection rate & cost• Low overall lifetime cost
SummarySummary
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
• Video detection works reliable.
• Video detection is the fastest way to detect.
• Video detection has the lowest false alarms rate.
• Video detection offers immediate verification via CCTV.
• AID, Automatic Incident detection is the best detection method for Incident management
ConclusionsConclusions
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Why use other incident detection if you will verify by video?
Why use other incident detection if you will verify by video?
Just use Video Incident detection
Directly
Thank You
Just use Video Incident detection
Directly
Thank You
THE Reference in Traffic Video DetectionTHE Reference in Traffic Video Detection
Tel Germany +49 (0) 5446 – 20 65 32E-mail: info@traficon.de
www.traficon.com
top related