Web Conference 2024.07.23 Curb
- Monthly on a Tuesday call at 9am PT, 12pm ET, 5/6pm CET
Meeting ID: 898 5980 7668 - Passcode 320307
https://us02web.zoom.us/meeting/register/tZ0lcuCgrjwsHNyZRagmc86b12iCmWGBHfjq
One tap mobile: +13126266799,,89859807668#,,,,*320307# US (New York)
Dial by phone: +1 929 436 2866 (US) (Find your local number)
Computer vision for curbside management
- Welcome (5 mins)
- Announcements - Mitch Vars (5 mins)
- Presentation
- Occupancy, courier pickup/drop-offs, and payment compliance, privacy, security - Umojo (10 mins)
- Computer vision elements in CDS Events
- Shareout and Discussion
- Turnstone
- SFMTA
- Curb Working Group Steering Committee members
- Working Group participants
- Hosts: Jacob Larson, Omaha
- Note Taker: Matt Warfield, Boston
- Facilitator: Mitch Vars, OMF
- Outreach: Michael Schnuerle, OMF
- Watch for Umojo issue around adding computer vision fields
- Register for the OMF Summit, free for members, and 50% off now with CoMotion early bird ticket
- 59 Attendees
- OMF Slides
-
Recording - Password
6fVVJNu@
-
Welcome (5 mins)
-
Announcements - Mitch Vars (5 mins)
- SMART Round 2 applications - EM, OMF will support any applicants who want to apply
- OMF Summit 2024 - November 12, 2024 in LA prior to CoMotionLA. Topics will include CDS, equity, access. Will be after the election and will be able to touch on political aspects of the future of work. Panels, workshops, opportunities to get your hands dirty. Connect with other folks in technology and transportation. Early bird promotion - 50% summit and CoMotionLA tickets. More information about speakers to come. If an OMF member you get free tickets.
-
Presentation
- occupancy, courier pickup/drop-offs, and payment compliance, privacy, security - Umojo (10 mins)
- Solar powered computer vision cameras, easy to install, low infrastructure
- Putting data into CDS format
- Recording information on vehicle activity - 30 second snapshots, not video- vehicle classification and color, occupancy, parking compliance, double parking, commercial parking (logos), traffic/bike/bus lane blocking. Coming soon is tracking scooter and bike parking.
- By end of year will have video platform that can record more information such as vehicle counts
- Batter system with solar panel, can last up to two weeks with 30 second intervals
- Standard height 16’
- A lot of spaces from a single view, one camera
- Cloud managed and controlled
- Privacy
- Fuzz out people, license plates
- Images stored based on city data retention policy
- Images stored in their infrastructure - FOIT not applicable
- Data analytics @ ground truthing
- Tracking data, dashboard, can provide any analytics someone asks about
- Ground truthing - checking how accurate AI is - video based
- Deployment
- Planned vs reality - permits, infrastructure challenges
- Questions
- Do they work at night - they can work at night but would require some more planning
- Do cameras spot animals - they might be able to but would need to be set up to do so
- occupancy, courier pickup/drop-offs, and payment compliance, privacy, security - Umojo (10 mins)
-
CDS events that can be shared out
- Existing optional elements
- Vehicle type, length, visible identifier, license plate number; linear curb location; lane type; some event purposes (here to drop off passenger, food delivery, utility work, construction); maybe propulsion types (electric, gas)
- Possible new optional elements
- Vehicle make, model, year, color, company, license plate state/province; confidence levels
- Umojo is creating GitHub on the subject and we can add elements we would like to see considered
- shared CDS Events draft additions sample fields Doc from Umojo
- Existing optional elements
-
Share out and Discussion
- SFMTA - been using LPR to manage meter spaces on blocks that use multi-space pay stations for about 3 years; expanded that technology to manage resident permit parking areas - no longer issue physical stickers, validate users by license plate number. By having these LPR cameras can leverage that technology to give a better insight on parking technology. SFMTA was a recipient of early valued pricing grants - on-demand pricing, curb management, sensors to understand real time parking data - linked to an app to see real time where parking was available, sensors did not last. Hesitant to go to a hardware solution given costs - using LPR data to get a snapshot of what is happening on blocks. Dedicated vehicle that is driving around the City surveying every meter block multiple times a day - on time intervals - to generate occupancy data. Conduent is LPR vendor. Now leveraging LPR reads from the parking enforcement team - as they are doing enforcement also capturing occupancy data, using anonymized plate reads. Anonymized feed that goes into a separate database from enforcement. Data is used as a coefficient to apply payment data to meters. Alex is streets data information manager - created algorithm that feeds data into. Shift towards paystation and pay by plate user adjustments - confused, not realizing they have to pay. Change in infrastructure of payment type has impacted payment compliance. Have not gone with fixed LPR for curb management. Stage 1 SMART grant, internal digital workflow tracking system - as making changes to the curb would be reflected in curb API. Looking to pull in data from private companies. Fixed cameras as part of initiative to improve public safety - are not using that data but supporting them on the installation. Will be installing fixed cameras for speed enforcement that will be LPR - they wont be directly utilized for curb management.
-
Things people would like to see in CDS that are there now, or could be there
- Have any other cities have positive experiences with using fixed cameras for speed management
- How long do you keep data from LPR cameras
- Have citywide data retention requirements, for occupancy data do not store data, de identify license plate data, it is only kept 24 hours to validate occupancy reads then it is dumped
- Enforcement data stored for 30 days
- Cameras are not providing video
