Skip to content

Latest commit

 

History

History
65 lines (35 loc) · 2.23 KB

File metadata and controls

65 lines (35 loc) · 2.23 KB

Cascaded-Lidar-Ground-Segmentation

build passing velodyne64 compliant

A C++ version for P. Narksri, E. Takeuchi, Y. Ninomiya, Y. Morales, N. Akai and N. Kawaguchi, "A Slope-robust Cascaded Ground Segmentation in 3D Point Cloud for Autonomous Vehicles," 2018 21st International Conference on Intelligent Transportation Systems (ITSC)

Introduction

This is an c++ version implementation on the paper "A Slope-robust Cascaded Ground Segmentation in 3D Point Cloud for Autonomous Vehicles".

The origin code is in Python, which is vesry time consuming for runnning one frame.

So I swtich the code to c++ version, but it still very time consuming that cannnot use in real time.

The code still have a lot of parts to improve, if you want to use it, you can change the code to speed up.

The workflow

Image text

How to use

Test platform

laptop:

the 7th i5
Ubuntu 16.04
pcl 1.8
OpenCV 3.3
CMake

build

mkdir build
cd build
cmake ..
make
./test1 [filename]

dataset

The data file is also from the paper https://bitbucket.org/n-patiphon/slope_robust_ground_seg

Some Result

1. transform the cloud to depth image

Image text

2. sections for second segmeantion

Image text

Image text

3. segmentation result

Image text

Image text