Project Objective
Drivooo is developed to assist drivers in real-time. As described earlier, it helps you to avoid distractions and prevent collisions. Moreover, it will utilize the camera of your phone to scan objects and keep drivers following in a lane and warn of potential crashes in real-time.
Socio-Economic Benefits
Road Accident is a global problem, in every 2 seconds, an accident happens. According to WHO Approximately 1.3 million people die each year as a result of road traffic crashes. Some modern cars provide the same features as drivooo but they are way too expensive to support all classes of society. Consequently, to overcome this problem, we are developing an AI app that will be installed on any android device free of cost to facilitate drivers in critical situations.
Project Methodology
Python
Step 1) Trained SSD Object Detection Model with over 8 classes and produces TFlite file.
Step 2) Implemented Distance Estimation using Focal Length Formula.
Step 3) Implemented Lane Detection Module
Java
Step 4) Load that TFlite file into Java Project.
Step 5) Implemented Distance Estimation by following steps as we've done in Python
Step 6) Also added, Distance estimation module by using libraries and following some same steps as we've done earlier in python.
Project Outcomes
A free app that is accessible to everyone and provide support in real-time, to almost every kind of car whether it’s modern or old. Mobile will be placed on the dashboard and requires a normal quality camera for image processing. The app is providing assistance in real-time by detecting objects and distance estimation module, whether the car is hitting the object or not. It will generate a voice alert when the user is about to collide with the car. Moreover, it also provides voice alert based upon lane departure.
