Detection of grass using laser scanner

Abstract

Smart vehicle is an rapidly increasing field. The ultimate goal is to develop the

ability for our cars to travel without human control. As a result, the necessity of

applying sensing system becomes more and more important. Laser range finder is a

kind of sensor which can provide real-time and accurate distance information.

In this dissertation, a single laser range finder is used to scan the surroundings.

These data are processed to produce the detection result. Main objective of the

work is to distinguish grass and non-grass and reflect the result in RVIZ. The main

algorithm is to do segmentation first, then use the points of each segment to fit a

line, and finally, calculate the covariance and compare the covariance to a threshold.