An effective method to segment vehicles and roads is proposed for autonomous vehicles using low-channel 3D lidar. The distance-view transformation is newly proposed to overcome the low density of top-view data of lidar. In addition, a dilated convolution structure is proposed to expand the receptive field of a convolutional neural network. The proposed network improves the accuracy of segmentation. The experimental results are presented to verify the usefulness of the proposed method.
Segmentation of Vehicles and Roads by a Low-Channel Lidar
IEEE Transactions on Intelligent Transportation Systems ; 20 , 11 ; 4251-4256
2019-11-01
2592623 byte
Article (Journal)
Electronic Resource
English
Engineering Index Backfile | 1910
|Safer roads with automated vehicles?
TIBKAT | 2018
|Heavy vehicles and light roads
Engineering Index Backfile | 1933
|