Self-driving cars using OpenCV

Abirami Ravishankar (~abirami-r)


6

Votes

Description:

An application of OpenCV can be used to aid self-driving cars. OpenCV is used to understand the nature of the obstacle present in front of the vehicle ie. by training it based on obstacles usually encountered on everyday Indian roads and the nature of the path to be taken. The car contains a camera that provides real-time input and identifies the object present in front of it. Based on this, it decides how much time to wait or to move around. Trained with several images from day-to-day scenarios of Indian roads gives less space for errors compared to humans driving/cars with inbuilt obstacle sensors. Existing applications of OpenCV in this regard are simply lane detection or image classifiers but have not been put together for an application like this. The maneuvering of the car would be based on the results of training per obstacle. It makes the car ‘smarter’ as it is trained and can take decisions based on the outcome of the algorithm. This revolutionary application would significantly reduce the number of accidents that occur every single day due to human error.

Prerequisites:

OpenCV, Python, a little information about self-driving cars

Video URL:

https://youtu.be/_niq-id0Esg

Speaker Info:

I am Abirami, a fourth-year Electronics and Communication Engineering student. I have been to the University of California, Berkeley for a semester abroad in my second semester. I am interested in Blockchain and IoT projects and have worked on over 10 projects over the last 4 years. I am experienced in building and working on projects in both fields.

Speaker Links:

https://github.com/abirami-r

Section: Embedded Python and IOT
Type: Talks
Target Audience: Intermediate
Last Updated: