Lack of real time flooding information in San Jose results in heavy damages and leads to loss of lives.

We are a Team consisting of 2 members. 

1. Deeya Viradia
2. David Doyoung Lee

1. Public awareness
2. State and city council member awareness

A website which will tell where the floods is. The red on the map suggest the area is flooded, orange is at risk.

For a Admin user they can send an alert to other agencies (see on the right )

1. Pitch https://www.goodwall.io/ideas/flood-finder-5f34
2. White paper https://dochub.com/deeya-viradia/xP5LzojKa1pz1M4K7ZDlO3/floodfinder-white-paper-11-15-pdf

final pitch: https://dochub.com/deeya-viradia/gYzM7mBwDjQ54mNKqJ6EoO/flood-finder-pitch-deeya-2-pdf?pg=6

Helped find a team. David and Deeya had pitched the idea to open 17. David with hardware experience

Deeya came with the AI experience.

@Deeya Viradia, One would assume you would get archived feed in the Open Data Portal of the SanJose govt, But I am not sure if you would find the archived videos there!? Traffic Authority? State Surveillance? Or check the city's data archives dept!

Looks like Sanjose is already discussing about geo-sensors in traffic lights to detect incoming emergeny vehicles and block traffic.

A PhD Candidate in ETH has been working on detectng the flood level from social media imagery. She had a good dataset of flood images, I am not sure if she tried with live traffic feed. She might have extended her study, her focus was an algorithm detecting water level from imagery.

https://arxiv.org/pdf/2007.06749.pdf

We can check with her availability and put you in touch if need be.

Coming to your crowdsourced situational awareness - which is also one of your approaches - check this https://www.giscorps.org/napsg_243/

• There are 22k+ street intersections in San Jose  and there are 7.5K+ traffic cameras (https://gisdata-csj.opendata.arcgis.com/datasets/street-intersections), which mean ~36% streets are covered with traffic cameras.
• San Jose city has a volunteer program to share homeowner's personal camera feeds to San Jose police department( https://www.sanjoseca.gov/your-government/city-directory/city-council/district-9/public-safety/d9-camera-rebate-program/frequently-asked-questions-faq-s). 45% of San Jose residents  have personal cameras on their property.
• To increase more coverage  we can install a flood sensors where there are no cameras. 

1. Flood Data Sensors (Open Source Hardware)

   • Placed on buildings to indicate flooding & water depth 

2. Camera Feed (Traffic Camera + Ring door bell camera from private citizens)( using ML/AI)

   • Open sourced AI algorithm to detect areas w/ floods & to gauge the depth of the water

3. Mobile Application (Crowdsourcing) 

   • Central platform for all data. Users can input locations that are flooded along w/ photos

Didn't find the archive data for traffic feeds

We need to scrape live traffic camera data and obtain historical flood models if it is available. 

Our solution is an alert system. So the system should send alert to Emergency personal and to the citizens. 

Crowdsourcing could be a phase II solution.  

Flood Sensor needs 3G capability so that it can constantly transmit flood data. After adding 3G capability we need to create an IoT server (AWS is preferred option). 

1st Generation Flood Sensor Developed. 

An Arduino and ultrasonic sensor were used for the flood sensor. The 1st generation FloodFinder sensor detects the presence of water and water levels. As mentioned in previous weeks adding 3G capabilities to the sensors will be the next steps for the 2nd generation flood sensor. The flood sensor is assembled using common tools and low-cost materials. 

Final Pitch

1. Provide short and long term goals.
2. Add traffic camera images
3. Make clear what crowdsourcing element
4. Timeline
5. Roadblock
6. Imapact