vipulkumbhar/aue893_perception_and_intelligence

This repository contains class-work and practice examples for the camera, RADRA, and LiDAR data processing for object detection. Deep learning methods are used for object detection from an image.

AuE8930_Perception_and_Intelligence

Intereseting problems from homework

1) Design and implement the approaches to find all park space frames with the four vertex points of each frame.( without deep learning algorithms)

1.1 Canny transform on gray scale image to extract out edges.

1.2 Hough transform to find lines in image and superimpose it on original image

1.3 Finding out continuous lines from various line segments and finding out end points

1.4 Line intersections (blue circles)

1.5 Mapping rectangles in parking space

2) LiDAR data processing

2.1 Point cloud view

2.2 Voxel filter (or box grid filter) to downsample all the 3D point cloud points to the 3D voxel space points

2.3 RANSAC algorithm to the 3D voxel space points to find a ground plane model

2.4 Remove all the ground planes points in the 3D voxel space points

2.5 Off ground points with ground plane points

2.6 Top view projection of point cloud data

2.7 Front view image with color based on depth i.e. distance from ego vehicle