Project information

Things I did!

  • Developed a car that uses PID control to navigate through a traffic environment simulated in Gazebo and ROS (as seen on the left)
  • Implemented edge detection, masking and perspective correction to get cropped license plates from camera feed
  • Constructed and trained an Optical Character Recognizer (OCR) model using a CNN architecture with “1 of k” encoding to read license plates using Tensorflow and Keras with 90% accuracy

Read more in the report (link above)