NightCrawler

Inspiration

Car accident statistics are jarring at night. Despite 60 percent less traffic on the roads, more than 40 percent of all fatal car accidents occur at night. Sunlight provides the strongest light source. When it goes down at night, there are a variety of man-made lights that help drivers safely navigate the roads. Our eyes will adjust to lower levels of light (pupils dilate in darkness and constrict in brightness), but they have difficulty functioning properly when switching from bright to dark, or vice-versa. This can happen quite a lot on the roads at night when you look directly into the headlights of oncoming traffic.

How I built it

The system relies on two apps and a drone. The mobile app ( Android ) connected to a Drone gets bird eye imagery from a camera attached on a VINCI drone. We use Facebook ( Graph Vision API ) to identify the street lights and give them IDs We use EmsembleJS library to identify the difference and yield the faulty disfuncted lights We use Here Maps API to help nighttime navigation to the faulty street lights

What it does

Using the drone nighttime imagery , the project makes use of imagery analysis to identify which street lights are not working at night. The system use Vicci drone imagery over time, runs the analysis, and produces a list of faulty street lights for the traffic manager. The goal is to have brighter and safer roads and highways.