Detecting lane lines for autonomous driving using a convolutional neural network

Summary: When driving on a highway, most steering is focused on centering the car between lane lines. For this task it is necessary to determine both the position and curvature of the closest pair of lane lines. This project uses a convolutional neural network to identify lane line pixels in video taken from a dashboard camera. We'll discuss the full process from preparing the training data to calculating the position and curvature of the lane lines. Full source code will be provided in Python/Keras.

Bio: Eric Lavigne is the Director of Software Development for MCNA Systems, which creates dental insurance software. Eric is studying machine learning, computer vision, and sensor fusion as part of Udacity's self-driving car engineer nanodegree.