The problem SafeRoute AI solves

According to the Road Accidents Records in India for the year 2021, there is a staggering annual number of road accidents, resulting in 18 fatalities per day. Notably, over 4.4 thousand accidents are attributed to potholes on roads. Potholes, beyond being a significant cause of accidents, are a persistent nuisance for commuters. In response to this, we have developed a pioneering solution utilizing a deep learning object tracking model. This technology identifies and tracks moving vehicles through dash cameras, concurrently marking them on a digital map for the benefit of other drivers. Additionally, our system issues friendly warnings to the driver in real-time and communicates data to responsible authorities, such as the Public Works Department and the Ministry of Transport. SafeRoute is thus positioned as a revolutionary approach to road safety, functioning as a guardian on the go.

Challenges we ran into

We hackers are fueled by challenges, driven to identify problems, apply their knowledge, and concurrently acquire new skills to address and solve issues. In the pursuit of resolving the problem statement at hand, our team encountered two primary challenges:

1. Integrating and Scaling the Model: While possessing a capable deep learning model for pothole detection is an achievement, seamlessly integrating and scaling it within a web environment presented considerable challenges.

2. Simulating Dash Cam Conditions: The absence of an actual dash cam necessitated the exploration of alternative methods to emulate similar conditions, introducing an additional layer of complexity to our approach.