Lidar Obstacle Detection

This is the first project of Udacity's Sensor Fusion Nanodegree program.

In this project I have implemented plane segmentation algoritm, KD-Tree and Eucledian clustering algorithm all from scratch as taught in the course lesson. The code might not be as elegant as a professional software engineer but the code is efficient and it runs as expected.

Building the code.

git clone
cd lidar_obstacle_detection
mkdir build && cd build
cmake ../
make

Once the make process is successful, an executable file under tha name lidar_obstacle_detection will be created.

You can run the code using the following command

./lidar_obstacle_detection

Usage

The code takes one or no command line arguments. The pcd data streams used in the project is present in the src/sensors/data/pcd dirctory. If you have your own data set that you world like to try, then put them in this directory path under a different directory name and you can provide dirctory name as a command line argument as shown in usage 2 below.

Usage 1: No command line argument

If you don't provide any command line argument, the code will take data_1 directory by default to run the code.

Command: ./lidar_obstacle_detection

Output: The result of running data_1 through lidar obstacle detection pipeline will be like the one shown below.

Usage 2: Directory name as command line argument

If you have a directory full of chronologically named pcd files, then you can put the files under the src/sensors/data/pcd directory and provide the directory name as command line argumet.

Command: ./lidar_obstacle_detection

For example, if to run the data_2 directory through the lidar obstacle detection pipeline you should use the following command ./lidar_obstacle_detection data_2

Output: The result of running data_2 through lidar obstacle detection pipeline will be like the one shown below.

Author: Brolin A

Email: abrolin56@gmail.com