khaamosh / Autonomous-Vechicle-Detection

This capstone project was created towards the fullfilment of the Computer Vision Course at UNSW

Autonomous-Vechicle-Detection

This capstone project was created towards the fullfilment of the Computer Vision Course at UNSW

Description

This capstone project is aimed towards the application of computer vision techniques aimed at aiding autonomous vechicles in the following:

• Estimation of Distance of the oncoming Vehicle
• Object Detection and recognition.
• Estimation of Velocity of the oncmoing Vechicle
• Lane Detection of the autonomous Vechicle

Tasks

• The first part of the project is aimed at the estimation of the objects in front of the car
  • For this task, there are two implementations
    • Vechicle Detection file
      • In this code, file, the classical computer vision techniques have been implemented to draw boxes around the vechicles.
  • Deep Learning implementaiton using yolov3 model.
    • Distance Velocity Estimation
• Estimation of Distance of the oncoming vechicles
  • We used our deep learning model using supervised learning approaches for estimation of distance
• Estimation of Velocity
  • We employed similar neural based approach for velocity estimation.