Autonomous Car:Deep Learning & Computer Vision for Beginners

Autonomous cars: Deep Learning and Computer Vision Using Python & OpenCV on Raspberry Pi

Created by: Haider Najeeb
Computer Vision Engineer

Created by: Muhammad Luqman
Heavy Roboticist

What Will I Learn?

• Build your own Self Driving Car (Mini-Tesla)
• Learn (Divide and Conquer Approach) of solving complex problems i.e. Detection -> (Localization + Classification)
• Segment Lane Lines using Computer Vision techniques i.e. Canny Edge detection & Color Thresholding
• Localize objects in images based on specific shapes using Algorithms i.e. ShapeApproxPoly and HoughCircles
• Estimate Straight Line Trajectories Using Hough Lines and Curved Lanes using Custom Algo.
• Learn about Artificial Neural Networks and why Convolutional Neural Network are best for classification in Images.
• Learn to identify the right Algorithm in OpenCV and how to tweak it to your requirements
• Build, Train and deploy Custom CNN model (Deep Learning) for classifying Signs.
• Profile/Time your program using cProfile in Python
• Compare S.O.A Tracking techniques available in OpenCV and identify most suitable for project requirements
• Optimize your Code using simple but very effective IP techniques and threading
• Make SDV Navigate Autonomously in Custom Track and also obey road speed limits
• Understand how to extract actionable data from images
• Gain all the knowledge required to enter to more advance versions of (SDV series) upcoming courses to come... 🙂
• Brief Overview of SSD for Sign Detection and why its not the solution for every Object Det. problem

Requirements

• Basic Python programming and OpenCV
• 3D model's STL files printing

Target audience

• Beginner Programmers looking for an exciting project in the upcoming domains
• Enthusiastic students wanting to get their feet wet in multidisciplinary projects
• Engineers wanting to embark in the fields of Computer Vision, Artificial Intelligence and Robotics
• Self Driving Cars Enthusiasts looking to build one of their own