README.md

Nvidia Board Setting

This document describes the dependency installation of nvidia embedded board.

Xavier-NX, Xavier-AGX and AGX Orin 32GB with JetPack5.0.2 are tested.

Requirements

JetPack 5.0.2 (Ubuntu 20.04)

Python 3.8.10, CUDA 11.4.14, CUDNN 8.4.1, TensorRT 8.4.1

Eigen 3.3.7, Ceres 1.14.0, Protobuf 3.8.0, NLOPT 2.4.2, G2O, OpenCV 4.5.5, PCL 1.9.1, GTSAM 4.0

1.Set default python version to python3.8

sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.8 1
sudo update-alternatives --install /usr/bin/pip pip /usr/bin/pip3 1

2. Apt install libs

sudo apt install libboost-dev libboost-filesystem-dev libboost-date-time-dev libboost-iostreams-dev
sudo apt install libopenmpi-dev libtbb-dev libopenblas-dev busybox cmake
sudo apt install gfortran libjpeg-dev libgeos-dev libturbojpeg clang-12 libelf-dev
sudo apt install libzmq3-dev gpsd libflann-dev libqhull-dev libgps-dev libgstrtspserver-1.0-dev
sudo apt install libeigen3-dev libceres-dev

sudo add-apt-repository ppa:borglab/gtsam-release-4.0
sudo apt install libgtsam-dev libgtsam-unstable-dev

3. Install python libs

sudo rm -rf /usr/lib/python3/dist-packages/numpy*
sudo rm -rf /usr/lib/python3/dist-packages/scipy*
sudo pip install Cython
sudo pip install numpy==1.23.4
sudo pip install flask flask_cors colorlog pycuda easydict pybind11 pyinstaller json-rpc
sudo pip install protobuf==3.16.0
sudo pip install scipy filterpy shapely python-can onnx gevent websockets zerorpc flask_socketio gevent-websocket

4. Install prebuilt libs

git clone https://github.com/w111liang222/LSD-Nvidia-Libs.git
cd LSD-Nvidia-Libs
sudo ./install.sh # need to run this script twice

5. Clone this repository and build the source code

git clone https://github.com/w111liang222/lidar-slam-detection.git
cd lidar-slam-detection/
unzip slam/data/ORBvoc.zip -d slam/data/

sudo tools/scripts/install_dependency.sh
sudo python setup.py install

TensorRT engine

If you want to run object detection on nvidia board, the pre-generated tensorrt engine files are provided.

You can download the trt files from the following links:

• Xavier NX Google Drive | 百度网盘(密码b75q)
• Xavier AGX Google Drive | 百度网盘(密码339b)
• AGX Orin Google Drive | 百度网盘(密码jir9)

Uncompress the zip file and move all *.trt files (detection_image.trt, detection_lidar_l.trt and detection_lidar_s.trt) to lidar-slam-detection/ sensor_inference/