We believe that people should be aware of real-time crime zones around cities before travelling to a certain place be it for work or a vacation. We want to make sure that everyone in the community stays safe and does not get caught in any violent accidents.
What it does
We use the NYU IT API to fetch past criminal data on the NYU campus. Our software gets the real-time locations of danger zones through image feed and updates the zones on a google map through heat-mapping. We created a database which stores coordinates of the danger zones and the activity which makes the zone unsafe. We used computer vision to detect the images and classifies them by returning tags as the output and thereby determining the probability of the of the situation being safe or unsafe. The web app combined with data visualization allows people to view real-time crime and danger zones in a particular place where an event took place.
How we built it
Have a local python script that takes screenshots at regular intervals from real time video footage from security cameras. We run openCV and google cloud vision on google compute engine to check for possible criminal threats with real-time processing. This information is updated on a database that keeps track of past and upcoming criminal attacks and is displayed to the user using heat-maps. The administrator has access to notify students on campus of possible threats, and keep them safe in the process. This information, and process is useful for new students, who are unaware, cautious, about new locations around the NYU campus.
Challenges we ran into
We faced a dilemma to run compute engine, and run an instance for the first time. Using NYU's api(MuleSoft) was a difficult task as well, and we could not go past the 502 bad gateway error at one point. We also experienced lags when running openCV script on our local machine.
Accomplishments that we're proud of
Inspite of these challenges we faced, we did not give up, and persistently tried to correct our mistakes. We integrated compute engine and NYU's api successfully, and are proud to have achieved this.
What we learned
We learned that despite our current limitations, there was also a way to overcome the problems we faced using different technologies. We learned to use compute engine, and understand its potential to create a powerful system.
What's next for SmartEye
Our current software is simply a prototype, working with basic functions, and targeting a specific sample audience. Due to security concerns, and time contrasts, our project is restricted by having only the current system to function. In our minds, for the future of SmartEye, we believe to implement and test our software with more efficient technologies, hoping the software would satisfy the general need of security everywhere.