Computer Vision, NMPC, HSV, Visual Path Following, Least Square Method, RANSAC.

The main goal of this work is to study and develop systems for path detection and pa- rameters extraction directly from the image plane, using computer vision techniques, to feed controllers to perform NMPC-based Visual Path Following. Detecting a path to perform image-based control requires efficient processing that can extract parameters pre- cisely even in the presence of inconveniences in the navigation environment, such as path failures, spots of lights or shadows, poor lighting conditions, noise in the image captu- red, and floors with non-homogeneous colors. Such problems compromise the controller’s effectiveness because they affect the way the parameters are extracted.