This is an assignment from Visual Perception for Self-Driving Cars course of Self-Driving Cars Specialization on Coursera.org.
In this assignment, you will:
- Use the output of semantic segmentation neural networks to implement drivable space estimation in 3D.
- Use the output of semantic segmentation neural networks to implement lane estimation.
- Use the output of semantic segmentation to filter errors in the output of 2D object detectors.
- Use the filtered 2D object detection results to determine how far obstacles are from the self-driving car.
numpy and opencv for python are installed.
xy_from_depth
: Estimating the x, y, and z coordinates of every pixel in the imageransac_plane_fit
: Estimating The Ground Plane Using RANSAC
estimate_lane_lines
: Estimating Lane Boundary Proposalsmerge_lane_lines
: Merging and Filtering Lane Lines
filter_detections_by_segmentation
: Filtering Out Unreliable Detectionsfind_min_distance_to_detection
: Estimating Minimum Distance To Impact
Bounding box and distance output:
Check out more details in Environment Perception For Self-Driving Cars - Learner - v1.ipynb