Abstract
For an automated driving system, accuracy in
detecting and recognizing the objects in the surrounding
environment is essential to ensure the safe driving. In
recent years, Light Detection And Ranging (LiDAR)
have been used as an external recognition sensor, which
plays a vital role in mapping, location, and recognition.
However, although LiDAR's 3D pointcloud information
is accurate, the scanned data are sparse in the long
distance. This paper proposes a sensor-fusion method
that using the optical flow method to obtain the range
information of feature points from straight line moving
objects and fuse them with pointcloud data. The results
show that our method effectively improved the density
of the 3D LiDAR pointcloud.
